Liquid-phase and gas-phase investigation of
biomolecules in a single experiment
†
H.-Y. Huang,aT.-J. Lo,aY.-C. Chen,*abJ.-B. Huaand P. L. Urban*ab
Interrogating three-dimensional structures of biomolecules is
instrumental to the fields of biochemistry and biophysics. Mass
spectrometry is a technique that can provide important data on the molecular weight of protein molecules in the gas phase. However, the relationship between the structure of biomolecules in the native
water-rich environment and the artificial gas-phase environment
cannot always be established with high certainty. Here we commu-nicate initial results obtained using a simple analytical approach which allows one to perform measurements of structural parameters of biomolecules in liquid phase and gas phase within a single experiment. Single-component samples are injected as plugs to a
microscale capillaryfilled with aqueous medium and driven
hydro-dynamically. As they traverse the capillary, they undergo zone broadening. The extent of this broadening is measured at two points along the capillary by UV absorption detection, and used to compute hydrodynamic radii according to the Taylor dispersion analysis theory. Further downstream, the dispersed zones leave the capillary
through the sharpened outlet aligned to the orifice of a mass
spectrometer. This multiple detection strategy allows one to confront protein conformation in the liquid phase (deduced from the hydrodynamic size) and the gas phase (deduced from mass spectral patterns).
Introduction
Water is the most wide-spread solvent of biomolecules occur-ring in nature.1Through billions of years, biomolecules evolved
their native functions being surrounded by the omnipresent molecules of water. However, water is not always the most desirable matrix for the studies of molecular structure and function. For example, analytical techniques such as
crystallography2 or mass spectrometry (MS)3 require the
samples to be delivered in solid or gaseous form, respectively. Due to their structural, energetic, and biocatalytic functions, proteins are essential to living organisms. MS is a powerful and versatile analytical tool with high specicity and sensitivity, and it has found applications in the study of protein structures.3For
example, electrospray ionization (ESI)-MS,4 nanoow
electro-spray ionization (nanoESI)-MS,5 and matrix-assisted laser
desorption/ionization (MALDI)-MS6have been used to analyze
proteins, and even study changes to their structures, or inter-actions with other biomolecules. Even though conformations of proteins can be affected by external factors such as solvent, temperature or pressure, mass spectra can provide a wealth of information on the structure of macromolecules. Results of such analyses are interpreted based on the assumption that the gas-phase conformations/functions of biomolecules reect the native ones which exist in aqueous environments.7,8 This
assumption needs to be veried upon case-by-case basis limiting the usability of these powerful analytical platforms in bioscience.
Taylor dispersion analysis (TDA) is a simple method that enables measurements of diffusion coefficients and hydrody-namic radii of molecules driven by laminarow.9–14TDA was
already used with MS to measure diffusion coefficients of ana-lytes.15 However, TDA is oen used in conjunction with UV
absorption detection, in particular with multi-point detection systems, which are currently available commercially. Notably, excellent compatibility of TDA with UV imaging detection has been demonstrated.16–18
Here we introduce a simple approach for investigating conformations of biomolecules in liquid and gas phase almost simultaneously. The method is based on the hyphenation of capillary-ow TDA – using UV imaging detection at two detec-tion windows on one capillary – with atmospheric pressure ionization mass spectrometry (Fig. 1). This simple setup enables measurement of hydrodynamic sizes of molecules by TDA performed right before the detection by mass spectrometry.
aDepartment of Applied Chemistry, National Chiao Tung University, Hsinchu 300,
Taiwan. E-mail: yuchie@mail.nctu.edu.tw; plurban@nctu.edu.tw
bInstitute of Molecular Science, National Chiao Tung University, Hsinchu 300, Taiwan
† Electronic supplementary information (ESI) available: Additional results and discussion, experimental details, and additional tables andgures. See DOI: 10.1039/c3ay41309h
Cite this: Anal. Methods, 2013, 5, 5908
Received 1st August 2013 Accepted 16th September 2013 DOI: 10.1039/c3ay41309h www.rsc.org/methods
5908 | Anal. Methods, 2013, 5, 5908–5911 This journal isª The Royal Society of Chemistry 2013
Analytical
Methods
COMMUNICATION
Published on 17 September 2013. Downloaded by National Chiao Tung University on 28/04/2014 01:49:34.
View Article Online
Results and discussion
In the proposed TDA-MS setup, setting up the ion emitter was straightforward since– unlike in conventional nanoESI – there was no need for establishing a direct electric contact with ground or potential bias.19,20 Table S1† summarizes the main
experi-mental parameters used in this study. As the nL volume sample plug traverses the capillary, it is recorded three times:rst, in two windows of a UV imaging detector, and, second, using a quad-rupole time-of-ight (Q-TOF) mass spectrometer with an ion source operating at atmospheric pressure (Fig. 1). Please note that this setup and the data treatment method described below are different from the ones implemented in a previous study,15in
which electrospray ionization mass spectrometry was used to monitor Taylor dispersion of an analyte. In the current study, UV absorption signals– representing analyte peaks at two detection windows– are used to compute the hydrodynamic sizes (and/or diffusion coefficients), while the MS signal is used to measure molecular weight and draw conclusions on the conformations of biomolecules in the gas phase.
Combined TDA and MS measurements
Performance of the proposed TDA-MS system wasrst exam-ined using test samples containing analytes with different molecular weights, including an amino acid (tryptophan, 105 M), a peptide (angiotensin II, 105 M), and a protein (myoglobin, 104M). Fig. 2 shows the TDA results (le) as well as mass spectra (right) of tryptophan, angiotensin II, and myoglobin, respectively. The data describing the peaks of the analyte zones traversing the two UV absorbance detection windows (t1and t2,s1ands2) were used in the determination of
hydrodynamic radii and diffusion coefficients. The parameters of the peaks recorded in the rst and the second detection window of the UV imaging detector were obtained by tting curves (e.g. Haarhoff–Van der Linde) to the raw data, and inputted to the following equation:21
Rh¼ 4kB T s2 2 s
12
phr2ðt2 t1Þ (1)
This way, the hydrodynamic radii of tryptophan, angiotensin II, and myoglobin were estimated to be 0.40, 0.48 and 2.1 nm, respectively. In eqn (1), the parameter kB is the
Boltzmann constant (1.38 1023Pa m3K1), T is the absolute temperature (293 K), h is the viscosity of medium (in this case, water; 0.0010 Pa s) while r is the radius of the capillary used (25 mm). It was pleasing to note that the measured hydrody-namic radii of the test analytes are similar to the values reported previously (tryptophan: 0.32 nm,22 angiotensin II: <1 nm,23
myoglobin:2.1 nm.24However, it should be noted that the
peak of myoglobin exhibited strong tailing. Therefore, decon-volution had to be conducted by manualtting. Incorrect tting led to erratic Rhvalues (Fig. S1†). The obtained diffusion
coef-cients of tryptophan, angiotensin II, and myoglobin were 5.4 1010 m2 s1, 4.5 1010 m2 s1, and 1.0 1010 m2 s1, respectively. As expected, the diffusion coefficient of myoglobin is in the order of the values reported for proteins.25
Aer passing through the two UV detection windows of the TDA system, the eluting analytes could also be detected using the Q-TOF mass spectrometer. One should note that the mass spectra obtained with this setup (Fig. 2) look similar to conventional ESI mass spectra (cf. ref. 4) because multiply charged ions derived from peptides and proteins can be
Fig. 1 Schematic diagram of the system for TDA combined with quadrupole time-of-flight (Q-TOF) mass spectrometry. The sample was injected hydrody-namically by pressurizing gas in the headspace of the sample vial, while the applied pressure was controlled with a digital manometer (left). Following the injection, the system was vented, and the vial with a mobile phase was mounted in the inlet. A UV imaging detector was used to record the separated zones traversing two detection windows along the fused-silica capillary column. The outlet of the capillary was set in front of the orifice of the mass spectrometer in order to enable transfer of molecules into gas phase and their ionization in a manner similar to nanoESI.
Fig. 2 TDA result plots and mass spectra obtained using the system shown in Fig. 1 after the optimization of relevant parameters: (A) tryptophan, (B) angio-tensin II, and (C) myoglobin. See the ESI† file for experimental details. Running buffer (20 mM ammonium acetate at pH 7.4) was pumped by applying a pressure of100 kPa (slight differences in migration times are due to the inability to control the pressure precisely).
This journal isª The Royal Society of Chemistry 2013 Anal. Methods, 2013, 5, 5908–5911 | 5909
Communication Analytical Methods
Published on 17 September 2013. Downloaded by National Chiao Tung University on 28/04/2014 01:49:34.
observed. Singly charged ions were recorded for tryptophan ([M + H]+at the m/z 205; Fig. 2A) while multiply charged ions
were observed when larger molecules, such as peptides (Fig. 2B) and proteins (Fig. 2C), were used as test analytes. Thus, the molecular weights of the analytes traversing the TDA column could be conrmed using MS. These results demonstrate that the on-line coupling of TDA and MS is useful for obtaining the information of hydrodynamic radii and molecular weights of analytes within a single run.
Inuence of system parameters on the mass spectra
In the following experiments we tested the inuence of various parameters on the assayed form of the analyzed protein in the gas phase. We hypothesized that the hydrodynamic pressure applied in TDA may affect the determination of hydrodynamic radii of the target analytes. Thus, TDA was further performed using tryptophan, platinum nanocluster, myoglobin, and oval-bumin as test samples and with different hydrodynamic pres-sures applied to the inlet vial (7–70 kPa). As the hydrodynamic pressure increased, the apparent hydrodynamic radii of proteins (myoglobin and ovalbumin) became smaller (Fig. S2†). On the other hand, the hydrodynamic radii of tryptophan and platinum nanoclusters (1 mg mL1) were unaffected by the change of hydrodynamic pressure. It is known that hydrody-namic pressure may have a signicant impact on protein structures.26–30 Protein unfolding (denatured) occurs when
exposed to high pressure. Various analytical methods have been used to study the structure of proteins when exposed to high hydrodynamic pressures. These methods include dynamic light scattering (DLS),31 infrared spectroscopy (IR),32 rapid mixing
methods,33Taylor dispersion analysis (TDA),12 and
high-pres-sure NMR spectroscopy.34While the inuence of pressure on
protein structure was apparent at higher pressures (700 kPa), here we observe a change of Rh when applying relatively low
pressures (7 kPa). We deduce the cause of the variation of apparent hydrodynamic radii of proteins in the current study is different from the ones discussed in the previous studies which tested the inuence of very high pressures ($200 MPa).35,36
These reasons for elevated hydrodynamic radius at low pres-sures (Fig. S2†) can be shear stress induced by the laminar ow in the capillary column, adsorption on the capillary wall, and possible curvetting bias. Further discussion on the inuence of hydrodynamic pressure and capillary geometry on the folding state is presented in the ESI† le.
It is generally known that the temperature of dry gas (N2)
released through the MS inlet may also affect protein confor-mation in the course of desolvation when using conventional ESI sources. Since the ionization method used in this study is not exactly the same as conventional ESI or nanoESI, but more similar to contactless atmospheric pressure ionization (C-API),20
it was appealing to examine the effects of dry gas temperature. Myoglobin was used as the test sample in this experiment since the molecular weights of the folded form (Mw¼ 17 566 Da) and
the unfolded form (Mw¼ 16 949 Da) of myoglobin are different:
this is because heme molecules are only retained in folded myoglobin.37Thus, according to the m/z distribution of peaks
corresponding to the multiply charged ions of myoglobin in mass spectra, the information about the folding state of myoglobin molecules can be obtained. Fig. 3 shows the mass spectra of myoglobin obtained at different dry gas temperatures (160, 180, 220, and 260C). The peaks at m/z 1952 (M9+) and 2196 (M8+) – observed at lower temperatures – correspond to multiply charged ions of folded myoglobin (Fig. 3A and B). However, at higher temperatures of dry gas, the spectra reveal the peaks at m/z 1884 (M9+) and 2119 (M8+), which correspond to the multiply charged ions of unfolded myoglobin (Fig. 3C and D). Thus, the structure of protein in the gas phase is inuenced by the temperature of the dry gas. If the information about the folding state of proteins is expected to be obtained, the use of a lower temperature of the dry gas (e.g. 180C) is recommended. The above results suggest that the current approach can be used to study protein conformation by obtaining hydrodynamic sizes and mass spectral proles in one run. They also show that the aspect ratio of capillary channels used for transferring protein samples, and as ion source emitters, as well as dry gas temperature, may inuence the mass spectral patterns, and even give rise to system-related artefacts.
Conclusions
In summary, we have proposed a fast, easy, and inexpensive approach to study biomolecules in the liquid phase and in the gas phase. Both datasets provide orthogonal and complemen-tary information for the studies in protein conrmations. This approach provides various benets: two pieces of information (hydrodynamic size and molecular weight) are obtained in one run of one analytical instrument, saving time and workload. The time gap between the measurement of hydrodynamic size and the measurement of molecular weight is decreased down to seconds, which may be important when studying some of the dynamic biomolecular systems. The total sample volume used
Fig. 3 TDA-mass spectra of myoglobin (104M) obtained at different dry gas temperatures: (A) 160C, (B) 180C, (C) 220C, and (D) 260C. See the ESI† file for experimental details. The running buffer (20 mM ammonium acetate at pH 7.4) was pumped by applying a pressure of100 kPa. The peak labels highlighted with blue colour font correspond to unfolded myoglobin, while the peak labels highlighted with red colour font correspond to folded myoglobin.
5910 | Anal. Methods, 2013, 5, 5908–5911 This journal isª The Royal Society of Chemistry 2013
Analytical Methods Communication
Published on 17 September 2013. Downloaded by National Chiao Tung University on 28/04/2014 01:49:34.
in analysis is very small (nanoliters). Coupling TDA to MS is straightforward, and it does not require utilization of additional accessories. Since this simple hyphenated system provides the values of hydrodynamic radii along with mass spectra, it may further be used to study the stability of folded structures of native and modied proteins. Furthermore, this approach can potentially be used while investigating non-covalent interac-tions between small molecules and their target proteins in liquid and gas phase, which would be useful in the study of therapeutic proteins with potential applications in biomedi-cine. To this end, there is a need for further validation studies involving the described hyphenation of TDA with MS. Testing many other kinds of biomolecules (amino acids and peptides) using the developed TDA-MS platform will certainly extend the practicality of this approach. It would also be appealing to couple the TDA-MS system with upstream separation tech-niques (e.g. nanoow liquid chromatography and capillary electrophoresis) to enable direct analysis of mixtures.
Acknowledgements
We thank the National Science Council of Taiwan for the nancial support, and Paraytec Ltd (York, UK) for providing us the ActiPix D100 UV imaging detector.
Notes and references
1 Y. Levy and J. N. Onuchic, Annu. Rev. Biophys. Biomol. Struct., 2006, 35, 389–415.
2 A. L. Spek, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2009, 65, 148–155.
3 B. Domon and R. Aebersold, Science, 2006, 312, 212–217. 4 J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong and
C. M. Whitehouse, Science, 1989, 246, 64–71.
5 M. S. Wilm and M. Mann, Int. J. Mass Spectrom. Ion Processes, 1994, 136, 167–180.
6 M. Karas, D. Bachmann and F. Hillenkamp, Anal. Chem., 1985, 57, 2935–2939.
7 G. Ben-Nissan and M. Sharon, Chem. Soc. Rev., 2011, 40, 3627–3637.
8 M. Sharon, J. Am. Soc. Mass Spectrom., 2010, 21, 487–500. 9 M. S. Bello, R. Rezzonico and P. G. Righetti, Science, 1994,
266, 773–776.
10 U. Sharma, N. J. Gleason and J. D. Carbeck, Anal. Chem., 2005, 77, 806–813.
11 H. Cottet, J.-P. Biron and M. Martin, Anal. Chem., 2007, 79, 9066–9073.
12 H. Cottet, M. Martin, A. Papillaud, E. Soua¨ıd, H. Collet and A. Commeyras, Biomacromolecules, 2007, 8, 3235–3243.
13 T. Le Saux and H. Cottet, Anal. Chem., 2008, 80, 1829–1832. 14 J. Østergaard and H. Jensen, Anal. Chem., 2009, 81, 8644–
8648.
15 S. M. Clark, D. G. Leaist and L. Konermann, Rapid Commun. Mass Spectrom., 2002, 16, 1454–1462.
16 A. J. S. Chapman and D. M. Goodall, Chromatogr. Today, 2008, 22–24.
17 D. M. Goodall, E. Bergstr¨om and M. Surve, Chromatogr. Today, 2010, 31–34.
18 Paraytec Application Note AN001: Measurement of hydrodynamic radius for a standard protein over a wide concentration range.
19 Y.-T. Wu and Y.-C. Chen, Anal. Chem., 2005, 77, 2071– 2077.
20 C.-H. Hsieh, C.-H. Chang, P. L. Urban and Y.-C. Chen, Anal. Chem., 2011, 83, 2866–2869.
21 Paraytec Technical Note TN001: Hydrodynamic radius using Taylor dispersion.
22 T. Simonson, C. F. Wong and A. T. Brunger, J. Phys. Chem. A, 1997, 101, 1935–1945.
23 C. Chatterjee, D. Martinez and J. T. Gerig, J. Phys. Chem. B, 2007, 111, 9355–9362.
24 Paraytec Application Note AN015: Precision and accuracy of protein size determination using the ActiPix TDA200 Nano-Sizing System.
25 O. Annunziata, D. Buzatu and J. G. Albright, Langmuir, 2005, 21, 12085–12089.
26 V. B. Galazka, E. Dickinson and D. A. Ledward, J. Sci. Food Agric., 2000, 80, 1219–1229.
27 G. N. Somero, Experientia, 1992, 48, 537–543.
28 K. Akasaka, H. Li, H. Yamada, R. Li, T. Thoresen and C. K. Woodward, Protein Sci., 1999, 8, 1946–1953.
29 V. V. Mozhaev, K. Heremans, J. Frank, P. Masson and C. Balny, Proteins: Struct., Funct., Bioinf., 1996, 24, 81–91. 30 D. A. Ledward, HighPressure Res., 2000, 19, 1–10.
31 B. Nystr¨om and J. Roots, J. Chem. Phys., 1983, 78, 2833–2837. 32 W. Dzwolak, M. Kato and Y. Taniguchi, Biochim. Biophys.
Acta, 2002, 1595, 131–144.
33 P. Stefanowicz, I. Petry-Podgorska, K. Kowalewska, L. Jaremko, M. Jaremko and Z. Szewczuk, Biosci. Rep., 2009, 30, 91–103.
34 J. Jonas, L. Ballard and D. Nash, Biophys. J., 1998, 75, 445– 452.
35 N. Hillson, J. N. Onuchic and A. E. Garcia, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 14848–14853.
36 J. N. Onuchic and P. G. Wolynes, Curr. Opin. Struct. Biol., 2004, 14, 70–75.
37 C. Moczygemba, J. Guidry and P. Wittung-Stafshede, FEBS Lett., 2000, 470, 203–206.
This journal isª The Royal Society of Chemistry 2013 Anal. Methods, 2013, 5, 5908–5911 | 5911
Communication Analytical Methods
Published on 17 September 2013. Downloaded by National Chiao Tung University on 28/04/2014 01:49:34.
