Can Protein Conformers Be Fractionated by Crystallization?

May 31, 2013 - ABSTRACT: Molecular crystallization typically singles out a specific conformation, or a set of conformations that are identical over la...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Can Protein Conformers Be Fractionated by Crystallization? Aoshuang Xu,†,§ Fenglei Li,†,§ Howard Robinson,‡ and Edward S. Yeung*,† †

Ames Laboratory-USDOE and Department of Chemistry, Iowa State University, Ames, Iowa 50011, United States Biology Department, 463, Brookhaven National Laboratory, Upton, New York 11973-5000, United States

‡

S Supporting Information *

ABSTRACT: Molecular crystallization typically singles out a specific conformation, or a set of conformations that are identical over large parts and may show some flexibility, from a mixture of equilibrating conformations in solution. To critically evaluate the selectivity of this process, human lactate dehydrogenase isozyme 1 (LDH-1) microcrystals were separately dissolved and subsequently assayed inside capillaries with electrophoretically mediated microanalysis (EMMA) at both the ensemble and the single-molecule level. While fragments from the same crystal exhibited identical enzyme activities, different crystals, even when grown from the same drop of mother liquor, showed markedly different activities. Activities of individual molecules from a crystal were found to be essentially identical, whereas molecules obtained directly from solution showed a 4-fold variation in activity. Furthermore, after storage at 37 °C, the distribution of single-molecule LDH activities from solutions of individual crystals broadened and approached that of LDH obtained from the original solution. X-ray crystallography also showed distinct conformations for single microcrystals and confirms that crystallization properly selects even small conformational variants of proteins and that the slow equilibration to multiple stable conformations in solution is responsible for the observed single-molecule heterogeneity.

T

solution so far have been missing. Second, most experiments have been carried out with commercially available or relatively crudely purified enzymes that may be heterogeneous in their primary structure, especially isoforms or post-translational modifications. Polakowski et al. showed that while molecules from highly purified bacterial alkaline phosphatase have identical activities, endogenous proteolysis results in heterogeneity.29 Other studies on enzymes that show static heterogeneity,30−32 however, have not been able to unambiguously confirm the role of conformational multiplicity. Capillary electrophoresis (CE) is a powerful technique for determining enzyme activities and has played an instrumental role in single-molecule enzyme studies.33−38 For singlemolecule studies,1 a solution containing much diluted enzyme solution and necessary substrate and buffer was introduced into a thin capillary and incubated for some time for products to accumulate before voltage is applied to drive the products past the detector. Because the enzyme concentration is extremely low, individual molecules are a few centimeters apart on average along the capillary. Local product pools from individual enzyme will not mix by diffusion during the incubation time and are detected as individual peaks. Those areas are proportional to the activities of the corresponding enzyme molecules. By using this approach, LDH-1 has been shown to exhibit static

raditional enzymology studies enzyme activity at the ensemble level to reveal the average catalytic behavior of the molecules. However, analogous studies at the single-enzyme level have revealed considerable molecule-to-molecule variations in terms of the activity.1−17 Furthermore, experiments following the reaction trajectories of single-enzyme molecules show that the reaction rate is constantly fluctuating.7−12,14,18−20 Enzyme catalytic activity is ultimately determined by the folded three-dimensional structure of the polypeptide.21 The inherent complexity of proteins gives enzyme molecules many structural degrees of freedom, leading to a rugged free-energy landscape and possible multiple conformational states at various local energy minima.22,23 Such systems provide a critical test of the inherent selectivity of the protein crystallization process. A fluctuating enzyme model where multiple conformations interconvert on the timescale of catalytic reaction was proposed,7 to explain the observed dynamic heterogeneity in enzymatic activity. This has been supported by conformational dynamics probed by electron transfer,24−26 florescence resonance energy transfer,9,27 and NMR28 experiments. Although conformational dynamics satisfactorily explains enzymatic dynamic heterogeneity, caution should be exercised in extending it to static heterogeneity that is relevant to protein crystallization. First, static heterogeneity was observed on a much longer timescale, usually tens of thousands of turnover cycles or more, pointing to much higher energy barriers between conformations such that conformers were stable over hours. Direct observations of such conformational states in © 2013 American Chemical Society

Received: March 13, 2013 Accepted: May 31, 2013 Published: May 31, 2013 6372

dx.doi.org/10.1021/ac400762x | Anal. Chem. 2013, 85, 6372−6377

Analytical Chemistry

Article

heterogeneity, a feature that was attributed to the simultaneous presence of several stable conformations in solution.1 X-ray crystallography studies (Tables S1 and S2 and Figure S1−S4 of the Supporting Information) revealed the identities of several conformations of LDH, where the active site loop of any of the four subunits can be in an open or a closed conformation.39 There can also be a mixed conformation (Figure S2 of the Supporting Information), where some subunits are in one conformation and the other subunits are in another conformation. Although the resolution is low, the distances between residues in the open and closed conformations are different beyond experimental uncertainties (Figures S3 and S4 of the Supporting Information). The direct isolation of multiple stable conformations of LDH by crystallization thus provides the opportunity to fully assess the contribution of structure to the static heterogeneity at the ensemble level and at the singlemolecule level.

different crystals appear to be random and are given in Table S2 of the Supporting Information. Activity Measurements. LDH was first dialyzed for at 4 °C for 36 h against 10 mM Tris-HCl and the 50 mM NaCl (pH 8.1) buffer, using the Slide-A-Lyzer 10000 nominal cutoff dialysis cassettes from Pierce Biotechnology (Rockford, IL). NADH (5 mM) was added to the dialyzed protein solution. The proteins were concentrated to ∼20 mg/mL (based on the final solution volume) at 4 °C, using 10000 nominal cutoff Microcon centrifugal filters from Millipore (Billerica, MA). Crystals were obtained at both 35% PEG 400 plus 8% PEG 4000 (v/v) and 34% PEG 400 plus 8% PEG 4000 solutions. For ensemble-averaged experiments, 12 crystals were harvested from four separate crystallization droplets of 35% PEG 400 and 8% PEG 4000 (v/v): two from droplet A, four from droplet B, six from droplet C, and two from droplet D. Seven crystals were harvested from one crystallization droplet E of 34% PEG 400 and 8% PEG 4000. The sizes of crystals ranged from 10 to 150 μm in each dimension, as estimated from the microscope image. Crystals were fished out of the droplets using Mounted CryoLoop (Hampton Research) of the appropriate size and immediately dissolved in a 6 μL (hLDH) dissolving buffer (1 mM sodium salicylate, 50 mM lithium L-lactate, 10 mM TrisHCl, pH 8.0). Each crystal from droplet D was carefully separated and split into pieces, and these pieces were dissolved and stored separately. For the single-molecule experiment, three crystals were obtained from one droplet F and one crystal from droplet G, both being droplets of 34% PEG 400 and 8% PEG 4000 (v/v). Crystals were dissolved in 6 μL of 20 mM Tris-HCl, pH 9.0 buffer. All dissolved crystals were frozen immediately at −80 °C until further manipulation. Ensemble-Averaged Enzyme Assay. To compare the relative enzyme activities across crystals, we designed a two-step capillary electrophoresis (CE)-based method. All CE experiments were performed on a Beckman Coulter ProteomeLab PA 800 CE instrument (Fullerton, CA) equipped with a UV absorbance detector filtered at 214 nm. Untreated capillaries with a total length of 60 cm (50 cm to the detection window), 75 μm i.d., and 365 μm o.d. (Polymicro, Phoenix, AZ) were used throughout. Enzyme solutions were first calibrated against an internal standard (1 mM sodium salicylate) by hydrodynamic injection at 0.2 psi for 30 s (Figure S5A of the Supporting Information). The enzyme solution was then diluted 5−20× with the CE running buffer (50 mM lithium L-lactate, 10 mM Tris-HCl, pH 8.0) before being injected for plug−plug mode EMMA (Figure S5B of the Supporting Information). The injection sequence for the assay is 0.1 psi, 10 s diluted enzyme solution; 2 kV, 99 s buffer; 1 kV, 30 s NAD+. Because NAD+ moves faster than LDH under this buffer condition, NADH was generated when the NAD+ zone passes the LDH zone during electrophoresis. The capillary cartridge was set at 25 °C and the sample tray was set at 6 °C throughout the experiments. The microliter-size droplets dry quickly once they are removed from the crystallization chamber and exposed to air. Breaking the crystals and picking up the fragments took up to a few min to complete. So, crystals a and b were first isolated into even smaller droplets to ensure that the fragments could be attributed to a given crystal. Single-Molecule Assay. A homemade CE instrument was used for the single-molecule assay. Briefly, 2.5 mW of a 364 nm laser was isolated from a multiline UV laser (I-90, Coherent, Santa Clara, CA) with an equilateral dispersing prism and focused with a 15× UV objective into an 11 μm i.d., 150 μm

■

EXPERIMENTAL SECTION Chemicals and Reagents. Purified human LDH isozyme 1 (LDH-1) was purchased from Calzyme (San Luis Obispo, CA). Lithium L-lactate, β-nicotinamide adenine dinucleotide (NAD +), its reduced disodium salt (NADH), and sodium salicylate were obtained from Sigma (St. Louis, MO). Crystallizationgrade tris(hydroxylmethyl)aminomethane (Tris), Tris hydrochloride, sodium chloride, sodium hydroxide, PEG 400, and PEG 4000 were purchased from Hampton Research (Aliso Viejo, CA). Protein Crystallization. LDH crystals were grown by the hanging drop vapor diffusion method at room temperature. For the X-ray studies, six crystals of hLDH-H4 (Human LDH-1) (a−f) were obtained by mixing equal volumes (2 μL each) of a solution comprising 20 mg/mL protein and 5 mM NADH in 10 mM Tris-HCl, 50 mM NaCl at pH 8.0, and a well solution containing 10% PEG 4000 and 30% PEG 400 in 0.20 M TrisHCl at pH 8.1. One crystal (g) was obtained by mixing equal volumes (2 μL each) of a solution comprising 20 mg/mL protein and 5 mM NADH in 10 mM Tris-HCl, 50 mM NaCl at pH 8.0, and a well solution containing 8% PEG 4000 and 30% PEG 400 in 0.20 M Tris-HCl at pH 8.1. Three crystals (A, B, and C) with the LDH/NAD binary complex were obtained in a similar fashion. Crystals were blocks or rods of dimensions ∼100 × 50 × 60 μm. Crystals were flash-cooled to liquid nitrogen temperature in their original crystallization mother liquor without other additives. X-ray diffraction experiments were done in a synchrotron beam at Brookhaven National Laboratory. The data were processed using HKL2000. All crystals were of P212121 symmetry and have identical unit cell dimensions. There is one homotetramer in the asymmetric unit. The crystal packing in all cases studied were identical, in part due to the arbitrary assignment of A through D in each crystal. This implies that the loop position as well as the cofactor type did not significantly perturb the protein scaffold. Structures were solved independently by molecular replacement with AMORE (Figure S1 of the Supporting Information). Graphical images of the structures were created by Pymol. The coordinates of a homotetramer of human LDH/NADH/ oxamate (PDB 1I0Z) were employed as a search model. The molecular models were improved by refinements with CNS and Xtalview. The summary of final refinement statistics is shown in Table S1 of the Supporting Information, and the different combinations of conformations for the tetrameric units in 6373

dx.doi.org/10.1021/ac400762x | Anal. Chem. 2013, 85, 6372−6377

Analytical Chemistry

Article

Figure 1. Ensemble assay of fragments from two crystals from the same mother liquor. (A) Polarizing micrograph of the two crystals. (B) Relative activity of individual crystal fragments. Column height is the average of three replicate runs. Error bars stand for 2σ.

Figure 2. (A) Relative activities (area of the NADH peak in Figure S5B of the Supporting Information) of 21 crystals obtained from five crystallization droplets. Droplets A, B, C, and D were from 35% PEG 400 and 8% PEG 4000 (v/v); droplet E was from 34% PEG 400 and 8% PEG 4000 (v/v). Column height is the average of three replicates, except for droplet D where all nine data from three crystal fragments are used for both crystals. Error bars stand for 2σ. (B) Histogram of relative activities of the crystals.

portions upon dissolving. The first portion from crystal 2 was analyzed on day 5, and the first portion from crystal 3 was analyzed on day 6. The other halves of the solutions of both crystals 2 and 3 were further split into equal portions upon day 9 and stored at −80 °C and at 37 °C, respectively. Both portions of crystal 2 were analyzed on day 17 and both portions of crystal 3 were analyzed on day 24. The solution of crystal 4 was split into 4 portions on day 25 and incubated at 37 °C for 0, 12, 24, and 48 h before storage at −80 °C again and analyzed on day 28. The samples sizes are listed in Table S3 of the Supporting Information.

o.d. fused-silica capillary (Polymicro). Fluorescence was collected at 90° by a 40× objective (Edmund Scientific, Barrington, NJ), through a 364 nm long-pass filter (Semrock, Rochester, NY) and two bandpass filters (430 to 630 nm and 420 to 520 nm) (Semrock), and focused onto a side-on photomultiplier tube (model R928, Hamamatsu, Bridgewater, NJ). The analog fluorescence signals were digitized by a pDaq55 A/D converter (Iotech, Cleveland, OH) and recorded at 10 Hz with a LabView program on a PC. A 36 cm section of the capillary was surrounded with a water jacket and temperature-regulated by circulating water. Two bath/circulators were individually maintained at 40 and 22 °C. Connections between the bath/circulators and the capillary jacket were achieved by coupling pairs (Colder Products, St. Paul, MN) so that water circulating in the capillary jacket could be instantaneously switched between 40 and 22 °C. Enzyme solution was diluted in steps of 103× or 102× to the desired concentration in 20 mM Tris-HCl, pH 9.0 buffer and stored on ice until the assay. The diluted enzyme solution (1 μL) was mixed with 1 mL of CE buffer and filled into the capillary electrophoretically at 21 kV at 22 °C for 7 min. The water bath was then switched to 40 °C for 1 h incubation. Product zones were driven through the detection window by 21 kV at 22 °C. Blank runs were performed between actual assays (Figure S6C of the Supporting Information). A standard solution of 2.0 × 10−9 M NADH (plug) was injected at the beginning of each day to verify the instrument performance and the peak plateau height was used as an external calibration for the single-enzyme product peaks. The solution of crystal 1 was analyzed on day 3 after storage at −80 °C. Both crystals 2 and 3 were split into two equal

■

RESULTS AND DISCUSSION An EMMA method with UV absorption detection at 214 nm was developed for comparing ensemble average activities of LDHs of different crystals. A plug of LDH solution was injected first, followed by a plug of running buffer, and then a plug of NAD+ solution. Because NAD+ migrates faster than LDH, the NAD+ zone catches up and passes the LDH zone before reaching the detector. With lactate present in the background electrolytes, NADH is generated upon contact of NAD+ and LDH. This assay strategy has several advantages. First, capillary electrophoresis directly samples the enzyme solution and calibrates its concentration via the UV absorption detector. Second, electrophoresis provides a small sampling volume so that only a few nanoliters of enzyme solution is used in each assay and multiple replicate runs can be carried out to evaluate the assay precision. Third, electrophoresis affords the separation of NAD+, NADH, and LDH so that they can be individually quantified with a single UV absorption detector. 6374

dx.doi.org/10.1021/ac400762x | Anal. Chem. 2013, 85, 6372−6377

Analytical Chemistry

Article

Ideally, both NADH and LDH would be quantified in a single electrophoresis run. However, the catalytic efficiency of LDH is so high (∼300 units/mg) that an LDH concentration that can be reliably quantified by the UV detector is too high to maintain a linear enzymatic reaction. A two-step method (Figure S5 of the Supporting Information) with sodium salicylate as the internal standard (IS) was devised to circumvent this problem. First, a large plug of LDH solution containing 1 mM IS was injected into the capillary so that the LDH concentration was first calibrated with IS by their relative peak areas. Then, the enzyme solution was diluted with the running buffer (without IS) and a small plug of that was injected for the activity assay. A diluted enzyme solution was used to keep the reaction in the linear region. Two crystals from droplet D (Figure 1A) were isolated and broken apart individually so that three pieces from each crystal (Figure 1B) were tested. Reproducibility better than a 2.0% relative standard deviation (RSD) was obtained on all six crystal fragments, with three parallel CE assays each. Activities of fragments from the same crystal showed 0.6% and 2.0% RSD for crystals a and b, respectively, confirming the homogeneity of the crystals. Breaking the crystals and picking up the fragments took up to a few minutes to complete. So, crystals a and b were first isolated into even smaller droplets before being broken to ensure the fragments isolated were from the same crystals. A student t test shows a significant difference (P < 0.05) between all nine measurements of activity of crystal a versus those of crystal b. Figure 2A plots the activities of all 21 crystals obtained from five droplets. The heterogeneity of LDH activity for different crystals is evident. Droplets A, B, C, and D are of the same initial condition [35% PEG 400 and 8% PEG 4000 (v/v)]; droplet E is 34% PEG 400 and 8% PEG 4000 (v/v). Although only a limited number of crystals were tested, this slightly different crystallization condition (droplet E) does not seem to affect the resulting activities (P = 0.12). We found that the crystal size does not correlate with either activity (correlation coefficient 0.07) or the precision of the assay (correlation coefficient 0.13). The histogram (Figure 2B) shows that the activity of single crystals is not continuous but rather falls into discrete levels. This correlates with X-ray crystallography results that crystallization isolates different conformations of LDH (Table S2 of the Supporting Information). Specifically, since each of the 4 subunits of the tetrameric protein can be in either the “open” or the “closed” conformation, there should be 5 different types of crystals and thus 5 different types of activities, assuming the two conformations exhibit different activities. The data in Figure 2B indeed depicts at least 3 distinct groups of crystal activities. Previous single-molecule experiments showed that LDH molecules from solution have different activities.1 Although slightly different experimental conditions were used here, a similar degree of heterogeneity was found for LDH molecules directly from solution (Figure S6A of the Supporting Information). In contrast, LDH molecules from the same crystal showed a much narrower activity distribution (Figure S6B of the Supporting Information) for all four crystals tested (Figure 3). The F test indicates that the activity variance is significantly different (P < 0.05) between LDH directly from solution and those from any one crystal source. The RSD of activities for LDH from the same crystal is about 10% or smaller in each case, whereas 26% is observed for LDH directly from solution. Judging from the S/N ratios, the precision of the

Figure 3. Activities (area of the NADH peak in Figure S6, panels A or B, of the Supporting Information) of single LDH molecules from different sources. Column height is the average of all measured molecules from the same source. See Table S3 of the Supporting Information for population size. Error bars stand for 2σ.

experiments is not expected to be worse than the 9% obtained on similar systems for repeated incubation of the same molecule.13 So, molecules from the same crystals are identical in activity within experimental error. Furthermore, the singlemolecule activities are significantly different among crystals (P < 0.05), except for crystal 1 and 4 (P = 0.26). This agrees with our ensemble average experiments (Figure 2B) that crystal activities seem to fall into a few distinct levels. Experiments were attempted in order to see whether the activity distribution of molecules from dissolved LDH crystals might change after storage at 37 °C. Figure 4A shows that after one week at 37 °C, the activities of LDH molecules from crystal 2 changed significantly (P < 0.05). It is worth noting that the other portion of crystal 2 stored at −80 °C showed only minor change (P = 0.65) from the portion assayed a week earlier. This shows not only that −80 °C storage preserves the enzyme but also that the differences in average activity observed between crystals are not artifacts caused by instrumental fluctuations. For crystal 3 (Figure 4B), although less significant (P = 0.17) compared to crystal 2, change was also observed after 2 weeks at 37 °C but much less change (P = 0.09) was observed after 2 weeks at −80 °C. The gradual broadening of activity distribution of crystal 4 was evident over the course of 48 h (Figure 4C). The average activity, however, stayed virtually unchanged. Figure 4D shows the overall activity changes of all three crystals. The changes in average activity were not uniform: crystal 2 decreased; crystal 3 increased, and 4 stayed virtually unchanged. The increase in the average activity of crystal 3 means that the observed changes were not caused by thermal or proteolytic damage to the LDH molecules. The broadening of the distributions (Figure 4) and the change in the average activities support the argument that multiple stable conformations exist in solution and that crystallization fractionates these conformations. A closer examination of the results reveals that both the average activities and their distributions for every crystal gradually approach those of the molecules from the original solution, suggesting equilibration among the different conformations of LDH molecules. The equilibration is apparently a slow process, and certain conformations, such as that of crystal 3, may be more stable than the others. For crystal 4, the half-life, as determined from the change in the width of the distribution over time, was 36 h. The heterogeneity of single-molecule activities of the original solution of LDH thus 6375

dx.doi.org/10.1021/ac400762x | Anal. Chem. 2013, 85, 6372−6377

Analytical Chemistry

Article

Figure 4. Activity changes on storage for solutions of single LDH molecules from three crystals. (A) Crystal 2 after 1 week at −80 °C and at 37 °C. (B) Crystal 3 after 2 weeks at −80 °C and at 37 °C. (C) Crystal 4 after 12, 24, and 48 h at 37 °C. (D) Crystals 2, 3, and 4, before and after storage at 37 °C vs those directly from solution. Column height is the average of all measured molecules from the same source. See Table S3 of the Supporting Information for population size. Error bars stand for 2σ.

Notes

arises from the complete equilibration among different conformations. In summary, electrophoresis-based methods were applied to compare the specific activities of LDH from different microcrystals. Ensemble average assay of LDH from crystals showed that they fall into distinct groups in terms of catalytic activity. Single-molecule activities are similar among all LDH molecules from a given crystal, attesting to their purity, but are significantly different from those obtained from the original solution, depicting selectivity of the crystallization process based on the slight differences in protein conformation. The distribution and activities of molecules from crystals eventually mimic that of the original solution after dissolution and storage at 37 °C. The experiments here demonstrate that conformational multiplicity exists in solution state and crystallization is able to fractionate these into distinct crystals, even when there are only minor differences in the conformation. Furthermore, equilibration among these conformations in solution over hours is responsible for the observed static heterogeneity of LDH activity.

■

The authors declare no competing financial interest.

ACKNOWLEDGMENTS

■

REFERENCES

E.S.Y. thanks the Robert Allen Wright Endowment for Excellence for support. We thank Dr. Mary Jo Schmerr for her valuable help in the ensemble assay study. The Ames Laboratory is operated by Iowa State University for the U.S. Department of Energy under Contract DE-AC02-07CH11358. This work was supported by the Director of Science, Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences, and Biosciences.

(1) Xue, Q.; Yeung, E. S. Nature (London) 1995, 373, 681. (2) Tan, W.; Yeung, E. S. Anal. Chem. 1997, 69, 4242. (3) Nichols, E. R.; Gavina, J. M. A.; McLeod, R. G.; Craig, D. B. Protein J. 2007, 26, 95. (4) Craig, D. B.; Arriaga, E.; Wong, J. C. Y.; Lu, H.; Dovichi, N. J. J. Am. Chem. Soc. 1996, 118, 5245. (5) Ashley, C.; Dyck, D. B. C. Luminescence 2002, 17, 15. (6) Craig, D. B.; Nachtigall, J. T.; Ash, H. L.; Shoemaker, G. K.; Dyck, A. C.; Wawrykow, T. M. J.; Gudbjartson, H. L. J. Protein Chem. 2003, 22, 555. (7) Lu, H. P.; Xun, L.; Xie, X. S. Science 1998, 282, 1877. (8) Edman, L.; Foldes-Papp, Z.; Wennmalm, S.; Rigler, R. Chem. Phys. 1999, 247, 11. (9) Zhuang, X.; Kim, H.; Pereira, M. J. B.; Babcock, H. P.; Walter, N. G.; Chu, S. Science 2002, 296, 1473. (10) van Oijen, A. M.; Blainey, P. C.; Crampton, D. J.; Richardson, C. C.; Ellenberger, T.; Xie, X. S. Science 2003, 301, 1235.

ASSOCIATED CONTENT

S Supporting Information *

This material is available free of charge via the Internet at http://pubs.acs.org.

■

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions §

Contributed equally to this work. 6376

dx.doi.org/10.1021/ac400762x | Anal. Chem. 2013, 85, 6372−6377

Analytical Chemistry

Article

(11) Velonia, K.; Flomenbom, O.; Loos, D.; Masuo, S.; Cotlet, M.; Engelborghs, Y.; Hofkens, J.; Rowan, A. E.; Klafter, J.; Nolte, R. J. M.; de Schryver, F. C. Angew. Chem., Int. Ed. 2005, 44, 560. (12) Flomenbom, O.; Velonia, K.; Loos, D.; Masuo, S.; Cotlet, M.; Engelborghs, Y.; Hofkens, J.; Rowan, A. E.; Nolte, R. J. M.; Van der Auweraer, M.; de Schryver, F. C.; Klafter, J. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 2368. (13) Shoemaker, G. K.; Juers, D. H.; Coombs, J. M. L.; Matthews, B. W.; Craig, D. B. Biochemistry 2003, 42, 1707. (14) Hatzakis, N. S.; Engelkamp, H.; Velonia, K.; Hofkens, J.; Christianen, P. C. M.; Svendsen, A.; Patkar, S. A.; Vind, J.; Maan, J. C.; Rowan, A. E.; Nolte, R. J. M. Chem. Commun. (Cambridge, U.K.) 2006, 2012. (15) Hsin, T. M.; Yeung, E. S. Angew. Chem., Int. Ed. 2007, 46, 8032. (16) Craig, D. B.; Chase, L. N. Anal. Chem. 2012, 84, 2044. (17) Mashanov, G. I., Batters, C., Eds.; Single Molecule Enzymology. In Methods in Molecular Biology, Vol. 778; Humana Press: New York, 2011; 778. (18) Choi, Y.; Moody, I. S.; Sims, P. C.; Hunt, S. R.; Corso, B. L.; Perez, I.; Weiss, G. A.; Collins, P. G. Science 2012, 335, 319. (19) He, Y.; Li, Y.; Mukherjee, S.; Wu, Y.; Yan, H.; Lu, H. P. J. Am. Chem. Soc. 2011, 133, 14389. (20) Velonia, K.; et al. Angew. Chem., Int. Ed. 2005, 44, 560. (21) Anfinsen, C. B. Science 1973, 181, 223. (22) Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598. (23) Piwonski, H. M.; Goomanovsky, M.; Bensimon, D.; Horovitz, A.; Haran, G. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, E1437. (24) Yang, H.; Luo, G.; Karnchanaphanurach, P.; Louie, T.-M.; Rech, I.; Cova, S.; Xun, L.; Xie, S. X. Science 2003, 302, 262. (25) Min, W.; Luo, G.; Cherayil, B. J.; Kou, S. C.; Xie, X. S. Phys. Rev. Lett. 2005, 94, 198302/1. (26) Li, C.-B.; Yang, H.; Komatsuzaki, T. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 536. (27) Kuznetsova, S.; Zauner, G.; Aartsma, T. J.; Engelkamp, H.; Hatzakis, N.; Rowan, A. E.; Nolte, R. J. M.; Christianen, P. C. M.; Canters, G. W. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 3250. (28) Vallurupalli, P.; Kay, L. E. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 11910. (29) Polakowski, R.; Craig, D. B.; Skelley, A.; Dovichi, N. J. J. Am. Chem. Soc. 2000, 122, 4853. (30) Juul, S.; Ho, Y.-P.; Koch, J.; Andersen, F. F.; Stougaard, M.; Leong, K. W.; Knudsen, B. R. ACS Nano 2011, 5, 8305. (31) Kang, S. H.; Lee, S.; Yeung, E. S. Angew. Chem., Int. Ed. 2010, 122, 2657. (32) Claessen, V. I.; Engelkamp, H.; Christianen, P. C. M.; Maan, J. C.; Nolte, R. J. M.; Blank, K.; Rowan, A. E. Annu. Rev. Anal. Chem. 2010, 3, 319. (33) Bao, J.; Regnier, F. E. J. Chromatogr., A 1992, 608, 217. (34) Regnier, F. E.; Patterson, D. H.; Harmon, B. J. TrAC, Trends Anal. Chem. 1995, 14, 177. (35) Bao, J. J.; Fujima, J. M.; Danielson, N. D. J. Chromatogr., B 1997, 699, 481. (36) Van Dyck, S.; Kaale, E.; Nováková, S.; Glatz, Z.; Hoogmartens, J.; Van Schepdael, A. Electrophoresis 2003, 24, 3868. (37) Novakova, S.; Van Dyck, S.; Van Schepdael, A.; Hoogmartens, J.; Glatz, Z. J. Chromatogr., A 2004, 1032, 173. (38) Glatz, Z. J. Chromatogr., B 2006, 841, 23. (39) Li, F. Automated High-Throughput Protein Crystallization Screening. Ph.D. Thesis, Iowa State University, 2006.

6377

dx.doi.org/10.1021/ac400762x | Anal. Chem. 2013, 85, 6372−6377