Anal. Chem. 2004, 76, 315-321
Direct Sampling from Muscle Cross Sections for Electrophoretic Analysis of Individual Mitochondria Hossein Ahmadzadeh,† Ryan D. Johnson,† LaDora Thompson,‡ and Edgar A. Arriaga*,†
Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, and Department of Physical Medicine and Rehabilitation, University of Minnesota, MMC 388, Minneapolis, Minnesota 55455
Muscle is a highly heterogeneous tissue. Practical approaches to sample selectively small regions of muscle cross sections would help to effectively utilize analytical techniques on muscle studies while taking into account tissue heterogeneity. In this report, semimembranosus muscle tissue cross sections were directly sampled and analyzed by capillary electrophoresis (CE) with laserinduced fluorescence detection (LIF). Prior to CE-LIF analysis, a small region in the muscle cross section was stained with 10-nonyl acridine orange (NAO) which is a mitochondrion-selective fluorescent probe known to form a stable complex with cardiolipin, a phospholipid found only in mitochondria. By micromanipulation, the injection end of the capillary was brought into contact with the tissue exhibiting fluorescently labeled mitochondria. Sampling from a region similar in size to the cross section of a single fiber was carried out by applying 11 kPa of negative pressure for 3 s. When an electric field of -200V/cm was applied, fluorescently labeled mitochondria electromigrated and were individually detected by postcolumn LIF detection. For each sample, the electropherogram displays a migration time window with a collection of narrow peaks. The collection of individual peak measurements is represented as a distribution of individual intensities related to cardiolipin content of mitochondria and a distribution of individual electrophoretic mobilities. Positioning the capillary injection end was sufficiently spatially accurate to deplete mitochondria in the sampled region upon repetitive injections. Treatment of a muscle cross section with a protease (trypsin) prior to mitochondria sampling resulted in a higher number of detected mitochondria, suggesting that one of the effects of this enzyme is a partial digestion of the muscles myofibrils, which eases the release of interfibrillar mitochondria entangled within these fibers. The protease treatment also resulted in changes to the electrophoretic mobility distribution of individual mitochondria, which may imply that partial digestion of proteins bound to the mitochondria contributes to the alteration in the electrophoretic mobility of mitochondria. The ability to sample a region as small as a single muscle fiber cross †
Department of Chemistry. Department of Physical Medicine and Rehabilitation. * To whom correspondence should be addressed. E-mail: chem.umn.edu. ‡
10.1021/ac034809g CCC: $27.50 Published on Web 12/11/2003
arrriaga@
© 2004 American Chemical Society
section and its direct CE-LIF analysis opens exciting possibilities for the direct analysis of muscle biopsies and mapping the mitochondrial electrophoretic properties in highly heterogeneous tissues. Skeletal muscle is a complex tissue in which thousands of parallel muscle fibers with irregular cross-sectional shapes run along the direction of muscle contraction. Each fiber contains an array of contractile myofibrils that is supported by an extensive sarcoplasmic reticulum, multiple nuclei, and mitochondria.1-3 Fiber function is also complex, as each fiber may be classified as type 1 or type 2, depending on whether it relies on oxidative phosphorylation or on glycolysis for ATP production.4,5 Therefore, as the proportion of each fiber type varies depending on the specific muscle type, physical training, disease, or aging, the architecture of a muscle is expected to vary accordingly. The chemical analysis of such complex tissues is challenging, because the outcome will depend largely on the sampling location, and averaging out subtle differences in a tissue is hard to avoid. One of the approaches to address these challenges is through the use of histological analyses in which micrometer-thick tissue slices are stained with specific markers for enzymatic activity or biochemical function. For example, staining of a muscle cross section with an ATPase marker and an NADH marker distinguishes between type 1 fibers and type 2 fibers with high and low mitochondrial content and ATPase activity, respectively.6,7 Using histological procedures, it has also been feasible to investigate muscle disease and aging.8 One of the described disease phenotypes is that of Ragged Red Fibers, which is characterized by a lack of cytochrome c oxidase activity and overexpression of succinate dehydrogenase activity in the mitochondria.9,10 Even when the ratio of Ragged Red Fibers to normal (1) McArdle, W. D.; Katch, F. I.; Katch, V. L. In Essentials of Exercise Physiology; Lea & Febiger, 1994; pp 298-307. (2) Cooper, G. M. In The Cell: A Molecular Approach; ASM Press: 1997; pp 423-466. (3) Marieb, E. N. In Human Anatomy and Physiology, 2nd ed.; The Benjamin/ Cummings Publishing Company, Inc.: Redwood City, CA, 1991; pp 246284. (4) Rambourg, A.; Segretain, D. Anat. Rec. 1980, 197, 33-48. (5) Thakar, J. H. Physiol. Chem. Phys. 1977, 9, 285-295. (6) Chayen, J. C.; Bitensky, L. Practical Histochemistry; John Wiley and Sons: West Sussex, England, 1991. (7) Dubowitz, V. Histological and Histochemical Stains and Reactions; Bailliere Tindall: London, 1985. (8) Lopez, M. E.; Van Zeeland, N. L.; Dahl, D. B.; Weindruch, R.; Aiken, J. M. Mutat. Res. 2000, 452, 123-138. (9) Scheffler, I. E. Mitochondrion 2001, 1, 3-31.
Analytical Chemistry, Vol. 76, No. 2, January 15, 2004 315
fibers is extremely low, histological analyses have made it possible to identify these few abnormal fibers, which would be unnoticed in common analytical procedures that require homogenization of the entire tissue. Direct analysis by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry is another analytical technique capable of analyzing small regions in a tissue, single neurons, and vesicles.11-18 When using this approach, the sample (or a blot) is transferred to the MALDI-target surface and treated with a matrix, and then the MALDI process is accomplished via a focused nitrogen laser, which scans the target to generate a mass spectrum map. Another technique with high spatial selectivity for tissue analysis is laser capture microdissection (LCM) that can selectively remove a small sample, (e.g., 3-µm diameter circle),19 from a tissue cross section while leaving the rest of the tissue section intact.11,20 In this technique, the tissue sample is covered with a thin ethylene vinyl acetate polymer film, an IR beam is temporarily focused to melt the polymer covering the desired region, and when the polymer solidifies, it entraps the biological material from the region of interest. The solidified polymer is then removed, and the embedded sample is analyzed by other techniques, such as polymerase chain reaction and mass spectrometry.11,21-25 LCM has also been used to sample cross sections from individual fibers from histological preparations and then used to study mitochondrial DNA deletions.26 Separation techniques, such as capillary electrophoresis (CE), have also been adapted to finely control the position of the capillary injection end and to better define the sampled region. Combining CE with microdialysis sampling has resulted in a powerful technique for continuous monitoring of small molecules, such as neurotransmitters and peptides, in extracellular fluid of a variety of tissues, both in vivo and in vitro.27,28 The analysis of cellular contents by introducing a single cell into the separation capillary (10) Lee, C. M.; Lopez, M. E.; Weindruch, R.; Aiken, J. M. Free Radical Biol. Med. 1998, 25, 964-972. (11) Xu, B. J.; Caprioli, R. M.; Sanders, M. E.; Jensen, R. A. J. Am. Soc. Mass Spectrom. 2002, 13, 1292-1297. (12) Chaurand, P.; Stoeckli, M.; Caprioli, R. M. Anal. Chem. 1999, 71, 52635270. (13) Garden, R. W.; Moroz, L. L.; Moroz, T. P.; Shippy, S. A.; Sweedler, J. V. J. Mass Spectrom. 1996, 31, 1126-1130. (14) Garden, R. W.; Sweedler, J. V. Anal. Chem. 2000, 72, 30-36. (15) Kruse, R. A.; Rubakhin, S. S.; Romanova, E. V.; Bohn, P. W.; Sweedler, J. V. J. Mass Spectrom. 2001, 36, 1317-1322. (16) Kruse, R.; Sweedler, J. V. J. Am. Soc. Mass Spectrom. 2003, 14, 752-759. (17) Li, L.; Garden, R. W.; Romanova, E. V.; Sweedler, J. V. Anal. Chem.1999, 71, 5451-5458. (18) Rubakhin, S. S.; Greenough, W. T.; Sweedler, J. V. Anal. Chem. 2003, 75, 5374-5380. (19) Bonner, R. F.; Emmert-Buck, M.; Cole, K.; Pohida, T.; Chuaqui, R.; Goldstein, S.; Liotta, L. A. Science 1997, 278, 1481-1483. (20) Chaurand, P.; Caprioli, R. M. Electrophoresis 2002, 23, 3125-3135. (21) Emmert-Buck, M. R.; Bonner, R. F.; Smith, P. D.; Chuaqui, R. F.; Zhuang, Z.; Goldstein, S. R.; Weiss, R. A.; Liotta, L. A. Science 1996, 274, 9981001. (22) Bhattacharya, S. H.; Gal, A. A.; Murray, K. K. J. Proteome Res. 2003, 2, 95-98. (23) Craven, R. A.; Totty, N.; Harnden, P.; Selby, P. J.; Banks, R. E. Am. J. Pathol. 2002, 160, 815-822. (24) Rekhter, M. D.; Chen, J. Cell Biochem. Biophys. 2001, 35, 103-113. (25) Tanji, N.; Ross, M. D.; Cara, A.; Markowitz, G. S.; Klotman, P. E.; D’Agati, V. D. Exp. Nephrol. 2001, 9, 229-234. (26) Aiken, J.; Bua, E.; Cao, Z.; Lopez, M.; Wanagat, J.; McKenzie, D.; McKiernan, S. Ann. N.Y. Acad. Sci. 2002, 959, 412-423. (27) Davies, M. I.; Cooper, J. D.; Desmond, S. S.; Lunte, C. E.; Lunte, S. M. Adv. Drug Delivery Rev. 2000, 45, 169-188.
316
Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
or monitoring the secretion of a single cell by CE also attest to the high spatial resolution sampling of CE.29-32 Given that the cross-sectional area of a 50-µm-i.d. capillary (∼2000 µm2) typically used in our CE separations is similar to a typical single skeletal muscle fiber cross section (∼3000 µm2), it is envisioned that direct sampling from a muscle cross section into a capillary may result in a spatially selective CE analysis. Previously, we have reported the analysis of mitochondria prepared in bulk using (CE) with postcolumn laser-induced fluorescence detection (LIF).33,34 The ability to detect individual mitochondria facilitated counting these organelles and describing distributions of their individual properties, such as cardiolipin content33 and electrophoretic mobilities.34 In this report, we describe the use of micromanipulation to control the position of the injection end of the separation capillary, sample mitochondria directly from a spot similar in size to a single muscle fiber cross section, and then analyze them by CE-LIF. Direct tissue sampling for CE analysis of organelles opens up exciting avenues to expand the use of this technique in combination with other histological preparations and in particular to study the properties of individual organelles in highly heterogeneous tissues. REAGENTS AND METHODS Reagents. Sucrose, DMSO, and ethanol were purchased from Fisher Scientific (Pittsburgh, PA). Trypsin and N-[2-hydroxyethyl]piperazine-N-[ethanesulfonic acid] (HEPES) were purchased from Sigma (St. Louis, MO). KOH was purchased from Aldrich (Milwaukee, WI). Fluorescein and 10-nonyl acridine orange (NAO) were purchased from Molecular Probes, Eugene, OR. CE buffer contained 10 mM HEPES and 250 mM sucrose, adjusted to pH 7.4 with KOH. All buffers were made using Milli-Q deionized water and filtered using 0.22-µm filters before being used. Stock solutions of 1 mM fluorescein and 1 mM NAO were made in ethanol and DMSO, respectively. Dilutions of these solutions were prepared immediately prior to use. Animal and Tissue Preparation. The Institutional Animal Care and Use Committee of the University of Minnesota approved the animal care protocol. The semimembranosus muscle was obtained from 30-month old Fisher 344 Brown Norway F1 hybrid rats (National Institute of Aging Rodent Colony). In general, the rat was weighed and then anesthetized with sodium pentobarbital delivered intraperitoneally (50 mg/kg body weight). Subsequently, the semimembranosus muscle was carefully isolated from the hind limbs, and the midbelly region of the muscle was divided into 4-6 parts. The preparations were anatomically orientated and mounted on a cork base (parallel fibers running perpendicular to the cork base) with tissue glue and rapidly frozen in isopentane cooled to -170 °C, precooled in liquid nitrogen for 10 s, and stored in airtight polyethylene bottles at -80 °C until sliced into serial (28) O’Shea, T. J.; Weber, P. L.; Bammel, B. P.; Lunte, C. E.; Lunte, S. M.; Smyth, M. R. J. Chromatogr. 1992, 608, 189-195. (29) Tong, W.; Yeung, E. S. J. Chromatogr., B 1996, 685, 35-40. (30) Yeung, E. S. J. Chromatogr., A 1999, 830, 243-262. (31) Krylov, S. N.; Arriaga, E.; Zhang, Z.; Chan, N. W. C.; Palcic, M. M.; Dovichi, N. J. J. Chromatogr., B 2000, 741, 31. (32) Stuart, J. N.; Zhang, X.; Jakubowski, J. A.; Romanova, E. V.; Sweedler, J. V. J. Neurochem. 2003, 84, 1358-1366. (33) Fuller, K. M.; Duffy, C. F.; Arriaga, E. A. Electrophoresis 2002, 23, 15711576. (34) Duffy, C. F.; Fuller, K. M.; Malvey, M. W.; O’Kennedy, R.; Arriaga, E. A. Anal. Chem. 2002, 74, 171-176.
cross sections. Serial cross sections (10 µm) were cut using a cryostat (Leica CM3050S, North Central Instruments, Plymouth, MN) maintained at -25 °C.35 The serial cross sections were mounted on silanized coated glass cover slips and stored at -20 °C until analyzed. The day of the analysis, the serial cross section was brought to room temperature and then either (i) directly NAO-labeled as described in the next section or (ii) treated with 5 µL of trypsin solution (0.05% trypsin in 53 mM EDTA.4Na) for 10 min. NAO Labeling. Two approaches to label the mitochondria in the muscle cross sections were used. In the first approach, using a micropipet, 5 µL of 5 µM NAO was delivered on top of the tissue cross section, and then the cover slip was stored in an airtight sealed container to prevent the sample from drying during the labeling reaction. Using this method, a larger area of the muscle cross section was fluorescently labeled. In the second method, a 20-µm capillary was mounted on a capillary holder controlled by a micromanipulator (Soma Scientific, Irvine, CA) and positioned to precisely deliver a microdroplet of 5 µM NAO to a specific region of the muscle cross section. This capillary holder has been described previously and is routinely used for single-cell analysis.36 The other end of the capillary was attached through a proper fitting to a syringe containing 5 µM NAO. The process was monitored with the inverted fluorescence microscope (Nikon Eclipse TE300, Fryer Co., Huntley, IL). When the capillary was positioned on the selected region, slight pressure on the syringe attached to the other side of the capillary delivered a microdroplet of the NAO to a very small spot on the tissue cross section. Upon reagent delivery and in order to prevent sample drying, the glass slide was removed from the microscope and stored in an airtight 200 mL Ziploc storage bag along with a small beaker containing 50 mL of CE buffer at 4 °C. The labeling reaction proceeded for 10 min, and then the muscle cross section was placed back under the microscope to proceed with a direct injection into the CE capillary. The fluorescently labeled region was easily identified when using the fluorescence mode of the microscope. Direct Tissue Sampling. The injection end of a 50-µm-i.d. capillary used for CE was mounted on the holder described above and brought into contact with the fluorescently labeled region of the tissue cross section by micromanipulation while observing under the inverted microscope. For controls, the same procedure was applied to an unlabeled region. Sampling was performed via a hydrodynamic injection that resulted from applying negative pressure for 3 s using an electronic 3-way solenoid valve (Parker Instrumentation, Fairfield, NJ). When open, the valve established a hydrodynamic connection with a reservoir whose height is 110 cm below the tissue cross section. Upon completion of the injection, the injection end of the capillary was removed from the holder and placed into the running buffer vial for CE-LIF analysis. For multiple injections from the same spot on the tissue cross section, the procedure described above was modified to maintain the capillary holder and the tissue cross section in the same position on the horizontal plane. To bring the capillary in contact with the tissue cross section, only the vertical position of the capillary holder was adjusted by micromanipulation. It was also necessary to add 5 µL of buffer to the tissue cross section and (35) Thompson, L. V. Aging Clin. Exp. Res. 1999, 11, 109-118. (36) Anderson, A. B.; Gergen, J.; Arriaga, E. A. J. Chromatogr., B 2002, 769, 97-106.
then cover it with a small beaker to prevent it from drying before subsequent injections. Capillary Electrophoresis. The custom-built CE-LIF system used in this work has been previously described.37 The 488-nm line from an argon-ion laser (Melles Griot, Irvine, CA) was the fluorescence excitation source. Emission in the range of 522-552 nm was spectrally selected using a 535DF35 interference filter (Omega Optical, Bratteboro, VT) and detected with a photomultiplier tube (Hamamatsu Corp., Bridgewater, NJ). To reduce scattering of the 488-nm excitation line caused by bubbles or particulates in the sample (including mitochondria) an additional rejection band filter (488-53D, OD4, Omega Optical) was placed in front of the interference filter. The photomultiplier tube output was electronically filtered (RC ) 0.01 s) and then digitized using a PCI-MIO-16E-50 I/O board controlled by Labview software (National Instruments, Austin, TX). The data were collected at 100 Hz and stored as binary files. Separations were performed using a 50-µm-i.d., 150-µm-o.d., poly(acrylaminopropanol) (poly-AAP)-coated capillary.38,39 The coating provides a hydrophilic surface that minimizes the adsorption onto the capillary wall, and it has been previously used in the CE-LIF analysis of mitochondria.34 Using the current monitoring method,40 the EOF of the coated capillary was measured to be 1.2 × 10-5 cm2 V-1 s-1. The optical alignment of the detector was optimized for maximum S/N ratio by using 1 nM fluorescein. The instrument variability to particle detection was determined by continuously electromigrating 1-µm-diameter fluorescently labeled polystyrene beads (Fluoresbrite, Polyscience, Inc., Warrington, PA) suspended in CE buffer. After a direct sampling from a tissue cross section, the separation was performed at -200 V/cm, and the capillary was reconditioned between runs by manually flushing for 2 min each using water, methanol, water, and CE buffer. Data Analysis. Igor Pro software (Wavemetrics, Lake Oswego, OR) was used for data analysis. Tabulation of peak intensities and migration times for individual peaks was performed using a custom-written Igor procedure, PickPeaks, that has been previously described.37 From the data tabulated by PickPeaks, the electrophoretic mobility of individual mitochondria was calculated. The program was set to select only the peaks with a threshold that is five times higher than the standard deviation of the background. This identifies those peaks corresponding to the migration time window of the electropherogram that are then used to calculate individual electrophoretic mobilities. Since these data sets fit poorly to parametric distributions (e.g., poor fit to multiple Gaussians), statistical comparison was done using the Kolmogorov-Smirnov test (KS test) that is suitable for nonparametric distributions.41 RESULTS AND DISCUSSION Direct Sampling from Muscle Tissue Cross Sections. Sectioning a muscle into 10-µm-thick cross sections gives access (37) Duffy, C. F.; Gafoor, S.; Richards, D. P.; Ahmadzadeh, H.; O’Kennedy, R.; Arriaga, E. A. Anal. Chem. 2001, 73, 1855-1861. (38) Ahmadzadeh, H.; Dovichi, N. J. Ph.D. Dissertation, University of Alberta, Edmonton, 2000. (39) Gelfi, C.; Curcio, M.; Righetti, P. G.; Sebastiano, R.; Citterio, A.; Ahmadzadeh, H.; Dovichi, N. J. Electrophoresis 1998, 19, 1677-1682. (40) Huang, X.; Gordon, M. J.; Zare, R. N. Anal. Chem. 1988, 60, 1837-1838. (41) http://www.physics.csbsju.edu/stats/KS-test.html.
Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
317
Figure 1. Sampling from a muscle cross section. A. Fluorescent region caused by labeling a semimembranosus muscle cross section with NAO that has been delivered via a microdroplet; labeling reaction proceeded for 10 min at 4 °C. B. The injection end of the capillary is brought in contact with the surface of the muscle cross section. The capillary causes changes in the appearance of the image, but the capillary lumen is clearly distinguishable. C. After sampling for 3 s with negative pressure, part of the fluorescence in the region disappears.
Figure 2. CE-LIF analysis of individual mitochondria directly sampled from a muscle cross section. Part A, electropherograms from an NAOlabeled cross section (trace 1), a nonlabeled cross section (trace 2), and NAO alone (trace 3). Part B, fluorescence intensity versus electrophoretic mobility for individual peaks with signal-to-noise > 5 shown in part A. Crosses correspond to NAO-labeled sample; squares correspond to a nonlabeled sample. Tissue cross sections were labeled as in Figure 1 and not treated with trypsin. Hydrodynamic injection was performed for 3 s, and CE separation voltage was -200 V/cm. The capillary was AAP-coated, and the electrophoretic mobility (µ) was calculated from the migration time (tmig) of each peak, capillary length (L) and applied voltage (V) using the formula µ ) L2/Vtmig.
to the interior of the muscle fibers, which then can be directly treated with the mitochondria-selective probe NAO. Figure 1A shows the cross section of a semimembranosus muscle from the Fisher 344 Brown Norway F1 hybrid rat that was treated with NAO. Besides the highly fluorescent region corresponding to the area where the probe was applied (4000 µm2), the irregular shape of the individual skeletal muscle fiber cross section is easily discernible under this fluorescent image. Although NAO is selective toward cardiolipin found in the mitochondrial inner membrane,34,42,43 in this image total fluorescence cannot be attributed to mitochondria because the unreacted NAO is contributing to the detected fluorescence, thus complicating the interpretation of the image. On the other hand, Figure 1A clearly displays the target area where fluorescently labeled mitochondria may be found. As described in the Experimental Section, the fluorescence of the target area facilitates micropositioning the injection end of the capillary (Figure 1B). This figure clearly shows that the capillary lumen is targeting an area that corresponds roughly to the cross section of one skeletal muscle fiber cross section. Changes in the illumination profile in the present arrangement preclude the specific target of one single skeletal muscle fiber and result in some imaging artifacts. This can be observed by the bright halo around the capillary perimeter (i.e., (42) Petit, J. M.; Maftah, A.; Ratinaud, M.-H.; Julien, R. Eur. J. Biochem. 1992, 209, 267-273. (43) Maftah, A.; Petit, J. M.; Ratinaud, M.-H.; Julien, R. Biochem. Biophys. Res. Commun. 1989, 164, 185-190.
318 Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
75 µm from the capillary center). Upon sampling by negative pressure, the change in intensity in Figure 1C with respect to Figure 1A-B indicates that fluorescently labeled material (i.e., mitochondria and free NAO) has been introduced into the capillary. In addition, Figure 1C shows that the fluorescent intensity is decreased, suggesting that not all the labeled material was taken up during the 3 s that the hydrodynamic injection lasted. Altogether and importantly, Figure 1 demonstrates that the fluorescent labeling of a tissue cross section with a mitochondria selective probe aids in selectively sampling fluorescently labeled material for CE-LIF analysis. CE-LIF Analysis of Individual Mitochondria. For CE-LIF analysis, following sampling from a specific fluorescent region in a muscle tissue cross section, the injection end of the capillary was placed in a vial containing CE buffer, and the separation was performed at -200 V cm-1. Figure 2A, trace 1 shows an electropherogram typical of the CE-LIF analysis of individual mitochondria. This electropherogram has 350 peaks with a signalto-noise ratio > 5 and peak width of 0.16 ( 0.02 s (average ( standard deviation) at 10% of the total peak height. A control experiment, performed under the same conditions without using NAO labeling (Figure 2A, trace 2), has 17 peaks and peak width of 0.20 ( 0.02 s. Fortunately, these peaks are only a small fraction (5%) of the total number of peaks. In addition to the low number of peaks, the peaks in this control also show low intensity, which is likely the result of residual scattering (or autofluorescence) of
other particulate material that was introduced during the direct sampling from the tissue cross section. Evaporation of the NAO labeling solution during handling may lead to precipitation of particulate material that could result also in the appearance of narrow detected peaks and bias the data interpretation. This possibility was ruled out by directly injecting the NAO solution from a microscope slide. As shown in the corresponding electropherogram (Figure 2A, trace 3) no narrow peaks were detected. Instead, a band (peak width at base, 25 s) corresponding to a fluorescent species that diffuses in solution during the separation time is observed. The combined data of Figure 2A and the selectivity of NAO toward cardiolipin found in the mitochondrial inner membrane42,44 suggest that individually detected peaks correspond to mitochondria sampled from the muscle cross section. Although peak height provides an idea of relative organelle size because it relates to the amount of NAO bound to each mitochondrion (and to a lesser extent to residual scattering), electrophoretic mobility is expected to be a complex function of the ζ potential, size, and morphology.45 These parameters have been investigated in bulk using models of similar dimensions, such as microspheres,46 liposomes,37,45,47,48 and nanotubes.49 Figure 2B further illustrates that fluorescence intensity and electrophoretic mobility for the detected peaks in the electropherogram (Figure 2A, trace 1) are not correlated. Therefore, the lack of correlation between peak height and electrophoretic mobility suggests that size is not a dominant factor in the electrophoretic behavior of muscle mitochondria. Overall, these results suggest that mitochondrial electrophoretic mobility and size are very heterogeneous within a sampled muscle cross section. Effect of Trypsin Treatment. In muscle tissue, there are two types of mitochondria: subsarcolemmal mitochondria (SSM) and interfibrillar mitochondria (IFM).50-52 While SSM are clustered beneath the sarcolemmal membrane of the muscle fibers and require gentle vortexing for extraction, IFM reside between the myofibrils of the muscle fiber and can be released after a brief exposure to a protease, such as trypsin or Nagarase.53 A comparison between trypsin treatment and no treatment prior to NAO labeling and CE-LIF analysis was carried out to also release IFM and increase the number of mitochondria per sample. The number of peaks detected per injection for trypsin treated and non treated muscle cross sections were 1800 ( 550 and 350 ( 50 (average ( standard deviation, n ) 3, Table 1). Thus, trypsin treatment likely partially digests the actin/myosin cytoskeleton filaments in the myofibrils (and probably anchoring proteins found on the mito(44) Keij, J. F.; Bell-Prince, C.; Steinkamp, J. A. Cytometry 2000, 39, 203-210. (45) Radko, S. P.; Stastna, M.; Chrambach, A. Electrophoresis 2000, 21, 35833592. (46) Duffy, C. F.; McEathron, A. A.; Arriaga, E. A. Electrophoresis 2002, 23, 2040-2047. (47) Radko, S. P.; Chrambach, A. J. Chromatogr., B 1999, 722, 1-10. (48) Radko, S. P.; Chrambach, A. Electrophoresis 2002, 23, 1957-1972. (49) Doorn, S. K.; Fields, R. E., 3rd; Hu, H.; Hamon, M. A.; Haddon, R. C.; Selegue, J. P.; Majidi, V. J. Am. Chem. Soc. 2002, 124, 3169-3174. (50) Hoppel, C. L.; Moghaddas, S.; Lesnefsky, E. J. Biogerontology 2002, 3, 4144. (51) Fannin, S. W.; Lesnefsky, E. J.; Slabe, T. J.; Hassan, M. O.; Hoppel, C. L. Arch. Biochem. Biophys. 1999, 372, 399-407. (52) Palmer, J. W.; Tandler, B.; Hoppel, C. L. J. Biol. Chem. 1977, 252, 87318739. (53) Manneschi, L.; Federico, A. J. Neurol. Sci. 1995, 128, 151-156.
Table 1. Effect of Trypsin Treatment on the Number of Mitochondrial Peaks no. of peaksa cross section no.
trypsin treatment
no trypsin treatment
1 2 3 mean RSD (%)
1782 2339 1285 1802 29
350 256 339 315 16
a Conditions for sampling and CE-LIF analysis are the same as for Figure 2.
chondrial surface),1-3 facilitating release of IFM mitochondria from the tissue cross section. As seen in Table 1, the relative standard deviation (RSD) for the number of peaks for the trypsin treatment and no treatment are 29 and 16% respectively (n ) 3). These RSD values are higher than those previously observed in the CE-LIF analysis of suspensions of latex microspheres of various dimensions, where RSD values varied from 3 to 15%.46 More variation in the number of peaks observed in the direct CE-LIF analysis of tissue cross sections is not surprising, because it is expected that this tissue is highly heterogeneous and that the number of mitochondria will depend on the fiber type being sampled. A comparison of the average electrophoretic mobility distributions for the individual data reported in Table 1 is shown in Figure 3A. To facilitate visual comparison, the y axis is offset for the trypsin-treated distribution, and both are normalized by the combined total number of peaks of three replicates. Using this representation facilitates a qualitative comparison between these two nonparametric distributions. However, a more rigorous statistical procedure for determining if two nonparametric distributions are different is the KS test. Using this test, the KS value between these two treatments is 0.000 037. Since the KS value is small (e.g., 0.05), which indicates that the two distributions are statistically equivalent.41 Further similarity between the two intensity distributions is appreciated by weighing each bin in the distribution with the corresponding fluorescence intensity value and adding up the corresponding products to calculate an overall weighted intensity. Using this approach, the weighted intensity for the peaks corresponding to the trypsin-treated and nontreated muscle cross sections are the same (0.927 and 0.931 V, respectively). This comparison suggests that the trypsin treatment has not affected the end result of fluorescent labeling of the muscle cross sections with NAO and that the mitochondria sampled from the muscle cross sections under each treatment (SSM and IFM for trypsin treatment) and SSM in the absence of trypsin treatment have similar fluorescence properties (i.e., similar cardiolipin content distributions). Sampling Mitochondria from the Same Region. To further assess the effectiveness of the process of sampling mitochondria directly from a muscle cross section, we performed consecutive injections from the same area in the fluorescently labeled cross section. As shown in Figure 4, in both trypsin-treated and nontreated cross sections, the number of detected mitochondria decreased with subsequent injections. For the trypsin treatment, the total number of peaks recorded for four consecutive injections from the same spot was 3259, whereas it was 524 for three consecutive injections from the same spot of a nontreated trypsin cross section. It can be observed that the ratio of the number of peaks for trypsin-treated over nontreated (5.7) is in agreement with the single sampling of different cross sections (6.2) that was described in the previous subsection. However, sampling from the same spot indicated that the sampled mitochondria depleted rapidly. In the last of four consecutive injections from the same 320
Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
Figure 4. Mitochondria sampling from the same cross section spot. The y axis represents the fraction of peaks with respect to the combined number of detected peaks from the consecutive injections at the same spot. For both the trypsin-treated (A series) and trypsinuntreated (B series), the peaks were counted from electropherograms. Electropherograms were obtained as described in Figure 2.
spot of trypsin-treated muscle, only 10% of the total peaks were sampled (Figure 4, point A4). Similarly, the last of three consecutive injections from the same spot of a trypsin-untreated muscle produced only 2% of the total number of peaks that were sampled (Figure 4, point B3). These results show that the number of detected mitochondria decreases upon consecutive sampling from the same spot, irrespective of trypsin treatment, and that trypsin treatment aids in releasing more mitochondria from a given muscle cross section. The electrophoretic mobility distribution for mitochondria from the same spot of muscle cross sections treated with trypsin and untreated is compared in Figure 5. For trypsin treatment (panels 5A), the cumulative distribution corresponding to consecutive sampling described in Figure 4 (A markers) resembles that one of Figure 3A, distribution 1. Similarly, for the nontreated cross section (panels 5B, 1-3), the cumulative distribution corresponding to consecutive sampling from the same spot in Figure 4 (B markers) resembles that one of Figure 3A, distribution 2. The data presented in Figures 4 and 5 illustrate that the injection end
Figure 5. Electrophoretic mobility distributions for mitochondria sampled from the same cross section spot. Parts A and B correspond to trypsin treatment and untreated muscle cross section data in Figure 4, respectively. Black histograms correspond to the first injections (A1, B1) in Figure 4. Histogram bins corresponding to subsequent injections are stacked on top of the previous ones (e.g., A4 on A3, A3 on A2, and A2 on A1).
of the capillary can be carefully positioned to sample mitochondria from the same region. CONCLUSIONS Sampling organelles directly from a single fiber-size region in a muscle cross section illustrates the compatibility of CE-LIF analysis with sampling from heterogeneous tissues. The findings reported here indicate that for a small muscle cross section region, (i) the sampled organelles are not dissolved if adequate CE separation conditions are selected, (ii) organelles can be counted, and (iii) the analysis produces distributions of individual electrophoretic mobilities and fluorescence intensities of individual organelles. This analytical concept may be further extended to obtain a CE-LIF map from a tissue cross section analogous to those maps obtained from direct tissue analysis by MALDIMS.15,20,54 When dissecting a muscle tissue, the histological map of adjacent cross sections in a series is easily matched. It is envisioned that direct tissue sampling by CE-LIF analysis could be guided with a histological map of an adjacent section. Relevant (54) Chaurand, P.; Schwartz, S. A.; Caprioli, R. M. Curr. Opin. Chem. Biol. 2002, 6, 676-681.
applications of this approach would be to investigate the number of mitochondria for a given fiber type or the properties of mitochondria in aged tissues containing Ragged Red Fibers. Merging histological approaches with CE-LIF may provide a powerful quantitative tool to better understand the role and properties of mitochondria present in different skeletal muscle fiber types and the alterations that they may experience in developmental stages, disease, and aging. ACKNOWLEDGMENT This work is supported by NIH R01-AG20866-01. We thank Janice Shoeman at the Department of Physical Medicine and Rehabilitation of the University of Minnesota for preparing the muscle cross sections and Sandford Weisberg at the School of Statistics of the University of Minnesota for helpful discussion on statistical analysis. E.A. is supported by 1K02-AG21453-01. L.T. is supported by NIH R01-AG17768 and NIH K02-AG21626. Received for review July 17, 2003. Accepted November 10, 2003. AC034809G
Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
321
