Novel mass spectrometry technology development for large organic particle analysis

Abstract

Recently, mass spectrometry has been extended to detect large organic particles and biomolecules, which include large protein polymers, organic polymers, nanoparticles, virions, microparticles and cells. Different novel technologies have been developed to detect these very large particles. In this review, a brief introduction to the technology development is given, and future perspectives on the applications are included. In terms of the detection of very large biomolecules, a macromolecular ion accelerator was developed to achieve sufficient molecular ion energy to reach the megavolt region and increase the detection efficiency. Ions with mass-to-charge ratios (m/z) reaching 30 000 000 were successfully detected. For larger organic or bioparticles, laser-induced acoustic desorption was developed for placing these particles inside a quadrupole ion trap. Measurements of the masses of mammalian and poultry erythrocytes, organic microparticles and cells were achieved by a mass spectrometer with laser-induced acoustic desorption, a frequency-scanning ion trap and charge detection. The mass distributions of these particles were also determined. For nanoparticles and viruses, the number of charges on each particle is too low for accurate determination by a charge detector. The direct detection of nanoparticles/virions by a charge amplification detector is not feasible due to the low velocities of these nanoparticles. A novel approach was developed based on the simultaneous measurement of the different sizes and different numbers of charges of each nanoparticle to derive the masses of the nanoparticles. Due to recent developments in the detection of large organic particles, mass spectrometry can be used to detect masses ranging from atoms to cells.

---

Introduction

Mass is an intrinsic value of atoms and molecules. The invention of the mass spectrograph by Aston¹ was a major scientific breakthrough. In early studies, mass measurements were confined to atoms and very small molecules due to fragmentation from electron impact ionization² and the limited sensitivity of photographic plate detectors.³ After several decades of improvements in mass spectrometry ionization technologies to obtain intact molecular ions and better detection efficiency, most organic compounds with masses less than 5000 Da can be detected. The main ionization methods used to obtain intact ions of large organic compounds (500–5000 Da) include chemical ionization,⁴ field desorption/ionization,⁵ plasma desorption⁶ and fast atom bombardment (FAB).⁷ Charge amplification detectors, including electromultipliers (EMs),⁸ channeltrons and microchannel plates (MCPs),⁹ made the detection of single small ions feasible. Nevertheless, the detection of biomolecules remains challenging. Following several decades of extensive effort,¹⁰ electrospray ionization (ESI)¹¹ and matrix-assisted laser desorption ionization (MALDI)^12,13 were developed in the late 1980s. John Fenn and Koichi Tanaka were honored with Nobel prizes in 2002 for their achievements. Since the invention of ESI and MALDI, mass spectrometry (MS) has become a key technology in high-throughput analysis of biomolecules. Proteomic, glycomic and metabolomic analyses, which have made a tremendous impact on biomedical research, are primarily based on mass spectrometry detection of biomolecules. In general, ions that are generated from ESI have a broad charge state distribution. Most of them have low m/z ratios, so they can be easily detected by EM or MCP. MALDI¹⁴ has been broadly adopted for the ionization of large molecules and is often coupled with a time-of-flight (TOF) mass spectrometer to detect large proteins.^15,16 Up to now, nearly all ion detection in commercial mass spectrometers is based on the ejection of secondary electrons.^17–19 For small incident ions (i.e., those with m/z ratios less than 100), the number of secondary electron produced is usually larger than one when the ion energy is 30 keV. The released secondary electrons are subsequently amplified in an electron amplification device, such as an EM, a channeltron or a MCP. These electron amplification detectors usually have a gain of 10⁶ to 10⁸ so that a single ion can be detected. Therefore, a single ion with low m/z can usually be detected. The efficiency of the ejection of secondary electrons is a strong function of the ion velocity. When the velocity of the ions is lower than the threshold energy, the efficiency of secondary electron ejection is too low to be detected.

A fixed ion energy is given for most singly charged ions in a regular commercial mass spectrometer with a time-of-flight analyzer. Most biomolecular ions produced by MALDI are singly charged. The velocity of an ion is proportional to the inverse of the square root of the mass. When the mass increases by 4 orders of magnitude, the velocity decreases by 2 orders of magnitude.^20,21 For an ion with an m/z of 10 000 000, the efficiency for secondary electron ejection can be much less than 0.001. This low efficiency implies that at least 1000 ions are needed to generate one secondary electron. The detection efficiency approaches zero when the velocity of an ion is significantly lower than 10⁴ m s⁻¹.²² Particle detection remains limited to those with m/z ratios less than 1 000 000. Reliable mass measurements of particles with masses greater than 1 MDa has been challenging.

The ability to detect bioparticles with masses exceeding 100 000 is valuable. For example, the detection of intact high-mass glycoprotein and detailed glycoproteomic analysis can reveal glycoprotein structures and help in our understanding of the biological functions. By crosslinking proteins to stabilize their tertiary and quaternary structures, mass spectrometry plays an important role in studying protein complexes, such as bovine prion protein (bPrP), and monoclonal antibody interactions. Mass spectrometry has become the instrument of choice for rapid characterization of antibody–antigen interactions²³ and biomolecular interactions. Mass spectrometry can also give the stoichiometry of glycoproteins.²⁴ Some protein aggregations have strong relationships with neurodegenerative disease, such as Alzheimer's disease (AD), Parkinson's disease (PD), Huntington's disease (HD), amyotrophic lateral sclerosis (ALS) and prions. The aggregates usually consist of fibers containing misfolded proteins with a β-sheet conformation.²⁵ From mass spectrometric analysis, the degree of aggregation can be calculated from the mass of the aggregated particles. DNA sequencing can also be achieved by using mass spectrometry.²⁶ MALDI-MS can characterize large non-covalent complexes, and laser-induced liquid beam ion desorption (LILBID), as well as non-MS-based techniques, permits analytical methods for studying protein complexes.²⁷ MS has been used to identify viral capsid proteins, detect viral mutants, characterize post-translational modifications, and measure intact viruses.²⁸ Nevertheless, large biomolecules must be efficiently detected to perform routine analysis for the aforementioned tasks.

To solve the detection limitation of 1 MDa, two types of detectors, cryogenic and inductive, have been developed to overcome the barrier for ions with high m/z ratios.^29,30 The detection of polystyrene and doubly charged polystyrene aggregates with masses reaching 4 MDa have been demonstrated by cryo-detection MALDI-TOF-MS.³¹ A cryogenic detector is based on the measurements of the released thermal energy by particle bombardment. This detector overcomes the difficulty of detecting ions with very large m/z. The sensitivity of a cryogenic detector is independent of the mass. Cryogenic detectors have been demonstrated to detect very large molecules, such as singly charged ions in the mega-dalton regime from immunoglobulin M and von Willebrand factor proteins, and viral capsids.^31–34 Nevertheless, the effective area of a typical cryogenic detector is much smaller than that of a typical MCP. Therefore, the overall detection efficiency is lower. In general, the physical size of the entire cryogenic device is large, and the cost of maintaining low temperatures is high.^37,38 The other high-mass particle detector is a specific charge detector developed by Hillenkamp and co-workers³⁵ for MALDI-TOF. This detector comprises a flat 18 mm diameter metal electrode as a Faraday charge collector followed by a fast, low-noise, charge-sensitive FET impedance converter/amplifier. The mass range was extended to 300 kDa³⁶ by lowering the electronic noise³⁹ with poor mass resolution. To improve the response time and sensitivity, macromolecular ion accelerators (MIAs) offer a novel approach by directly enhancing the ion energy. The high velocity of the charged particles can be achieved by stepwise acceleration, leading to improvements in the detection efficiency. Fig. 1 shows the mass distributions of various particles. It is clear that the masses of many particles, such as chromosomes, very large organic polymers, nanoparticles, viruses, microparticles and cells, are heavier than 1 mega-dalton. The limitation of MIA is that the number of accelerated electrodes is not unlimited due to the velocity spread of the desorbed biomolecular ions. The detection efficiency decreases fast in the mass range of nano- and micro-particles. Nevertheless, the masses of nano- and microparticles are still too low to be measured by a microbalance. Therefore, an important goal is to measure particles with masses greater than mega-daltons but less than one microgram.

Fig. 1 The full range of mass distributions that can be measured by mass spectrometry.

Nanoparticles serve important roles in several fields. Hazardous air pollutants, such as fine particulates with sizes less than 2.5 micrometers, can be inhaled and can penetrate into the lungs to induce inflammation and cardiovascular diseases.⁴⁰ Other harmful nano-size infectious pathogens, such as viruses, will infect humans to cause influenza, chickenpox, herpes, HIV, SARS, etc. The rapid detection of pathogens has become important in recent years. For drug delivery systems, liposomes are being developed to serve as nanosized drug carriers.⁴¹ The quality control of liposomes is critical to assure the quantity of drug being delivered. Measuring the mass and size distribution of nanoparticles is difficult. The assurance of the size distribution of nanoparticles during synthesis is essential in achieving high quality control.

Some micro-bioparticles, such as cancer cells and bacteria, play key roles in human health. A rapid method to detect various cells and bacteria is in critical demand. Mass spectrometry can serve this purpose. The nano- and micro-particles analyses developed in the last decade include a nanoscale cantilever^42,43 based on changes in the resonant frequency as a function of the virus particle mass binding to the cantilever surface. A quartz crystal microbalance (QCM)^44–47 consists of a disk of crystalline quartz with gold electrodes on the upper and lower surfaces. The crystal quartz is typically induced to undergo a shearing oscillation due to the converse piezoelectric effect by the application of an alternating voltage to the electrodes. Charge-reduced electrospray size spectrometry⁴⁸ was used for the analysis of bacteriophage viruses with molecular masses ranging from 3.6 MDa up to the gigadalton range. Discrete conductance changes due to binding and unbinding by a single virus can be detected by an electrical detector⁴⁹ using nanowire field effect transistors. For gas-phase electrophoretic mobility molecular analysis (GEMMA),^50,51 protein particles representing a range of sizes can be separated by their electrophoretic mobilities. A charge-detection mass spectrometer^52–55 could also be used to monitor biologically important interactions, such as DNA – protein, virus – antibody, and virus – cell receptor complexes. Microscopy-based mass spectrometry⁵⁶ is capable of determining the absolute mass of a single whole virus with a size as small as 80 nm. Nevertheless, these methods have not achieved rapid and convenient detection at the level of a single nanoparticle/virus. They often require relatively complex sample preparation. Viruses have to bind to specific compounds on the surface of the detector. Considering antibodies as an example, the quantity of the antigen can be estimated by their binding to an antibody conjugated to a nanoparticle. Up to now, there are only a few experimental reports on the mass distribution of nano- and micro-particles. Mass spectrometry is a unique method for the analysis of unknown biological particles.⁴⁸

It is not easy to place Nano- and micro-particles into the gas phase for mass spectrometry analysis, even with MALDI. Laser-induced acoustic desorption (LIAD) was developed to desorb particles.^57–62 Recently, LIAD was successfully used to desorb viruses and nanoparticles into the gas phase for mass analysis.⁵⁶ A quadrupole ion trap with a fixed voltage and scanning frequency was used to obtain m/z ratios for nano- and micro-particles. The secondary electron ejection efficiencies are too low to detect large bioparticles with m/z ratios higher than 10⁸. A charge detector instead of a charge amplification detector was used. This experiment demonstrated that the mass region for mass spectrometry can be extended to detect nanoparticles using a charge detector when an individual nanoparticle or virus is trapped inside the QIT. There are two parameters, namely, m/z and the number of charges (z), that must be measured to obtain the mass. However, the number of charges for each small nanoparticle (<100 nm) is typically less than 100, and the electronic background is typically equivalent to ∼500 charges or higher. Therefore, a precise measurement for each nanoparticle is extremely difficult if not impossible. Nevertheless, the advantage of ion trap is the ability to accumulate over 10⁵ charges after a few laser shots. Thousands of singly charged or low charged nanoparticles can be accumulated within the trap to surmount high electronic backgrounds. To reduce electronic noise, the final spectra were generated from the sum of several spectra similar to ESI for biomolecular ions. The masses of the nanoparticles were obtained by de-convolution of the charge states of the particles or cluster numbers. The characteristics of nanoparticles with a size greater than 200 nm were similar to those of microparticles, which typically have charge numbers higher than one thousand for a single nanoparticle. The mass analysis for nanosized particles can be enhanced by charge detection of the particles.

LIAD was applied to desorb cells/microparticles and/or virus/nanoparticles without the need of a matrix to prevent any interference from single large particles originating from the matrix molecules produced by MALDI.^63–66 It is not necessary to produce protonation of the cell/microparticle samples. Most desorbed cells/microparticles are naturally charged. The detailed mechanism is still not well understood. In general, these desorbed microparticles often carry a large number of charges for each single microparticle. Most cells or microparticles within a vacuum chamber were found to have more than 1000 charges. Therefore, a charge detector can be used to measure the number of charges for each ejected single cell or microparticle directly.⁶⁷ Because the m/z ratio can be determined for a cell/microparticle using an ion trap mass spectrometer and the charge detector can be used to determine the value of z for each single particle, the mass of a cell or microparticle can be obtained. Microscopy-based mass spectrometry was used with light-scattering measurements to determine the m/z ratios for these trapping bioparticles.⁵⁶ To determine the masses of these microparticles, the number of charges of the trapping microparticles have to be changed by electron bombardment based on the observation of the trapping pattern.⁶⁸ Due to the low cross section of electron bombardment to change the number of electrons on the microparticle, this process is time-consuming. Overall, approximately 30 minutes is needed to determine the mass of one trapping microparticle. The mass distributions of most bioparticles are expected to be broad. To obtain the mass distribution, many single microparticles must be measured to obtain the mass distribution. This method becomes impractical for performing routine mass distribution measurements of microparticles.

When the measurements of nanoparticles/viruses and microparticles/cells are successfully achieved, mass spectrometry will have the capability to measure particles from atoms to cells, namely, the mass regime from a few daltons to 10¹⁷ Da. In this article, we review various mass spectrometry related technologies for the measurement of these large particles.

Macromolecular ion accelerator

The experimental schematic of a macromolecular ion accelerator (MIA) is shown in Fig. 2. The accelerator consists of a series of plates with equal spacing between adjacent ones. These plates are powered by time-regulated pulsed voltages to achieve acceleration. All of the pulses are edited and preset by an arbitrary waveform generator. When the desorbed ions pass a specific electrode, a pulsed voltage is applied to conduct acceleration. Large molecular ions are produced by MALDI with the sample probe mounted on the first plate. Programmable pulses consisting of a sequence of time-varying fields between two adjacent electrodes are applied to increase the kinetic energy of the selected ions. The pulse sequences are fine-tuned to optimize the signals detected by the MCP detector.

Fig. 2 Macromolecular ion accelerator. The probe is mounted on the first plate with an applied DC voltage of 25 kV, and the sample is exposed to the third harmonic of the Nd:YAG laser (355 nm) for desorption/ionization. The 2nd, 5th, 8th… plates (blue) are wired together and connected to the first switch (SW1). The second switch (SW2) and third switch (SW3) are connected to two other sets of electrodes, which are represented in green and turquoise, respectively.

MIA-accelerated macromolecular ions with masses up to 30 MDa have been pursued. With an acceleration energy of 200 kV, IgG (molecular weight ∼150 kDa) and fibrinogen (molecular weight ∼340 kDa) (ref. 69) ions with a single charge were detected, and the results show a signal increase versus the increase of ion energy (Fig. 3). Good agreement was obtained between the calculated results for the terminal velocities and the experimental measurements of the arrival times. This finding verified the efficiency of the acceleration.

Fig. 3 The temporal evolution of the effect of acceleration. (a) The acceleration of IgG with various acceleration energies from 25 kV to 205 kV. (b) The acceleration of fibrinogen with energies from 55 kV to 205 kV.

The agglomeration of all-trans retinoic acid was generated without the need for a special sample treatment. Fig. 4 shows the successful acceleration for 0.5 MDa, 1 MDa, 5 MDa, 10 MDa, 20 MDa and 30 MDa. For each molecular weight, programming of the pulsed acceleration in advance can be used for ions with specific m/z ratios. These results demonstrate that MIA can be used for the analysis of very heavy ions. With molecular ion acceleration, the first detection of ions with m/z ratios reaching 30 000 000 with an electron amplification device was reported. Thus, the acceleration and detection of molecular ions up to 30 MDa can be a milestone in the development of accelerators and in mass spectrometry detection technology.

Fig. 4 The acceleration of polymeric ions formed by all-trans retinoic acid at 500 kV. Due to its self-polymerization, all-trans retinoic acid can naturally form large polymers: (a) 0.5 MDa, (b) 1 MDa, (c) 5 MDa, (d) 10 MDa, (e) 20 MDa and (f) 30 MDa. All of the polymers were singly charged and created in the MALDI process.

Laser-induced acoustic desorption

The schematic of the LIAD-MS is shown in Fig. 5. Laser-induced acoustic desorption (LIAD) was originally developed for biomolecule detection without the need of a matrix.^58,59,70 LIAD successfully desorbed Al₂O₃ microparticles (Fig. 6). These results indicated that LIAD can successfully desorb microparticles. The trajectories of the desorbed particles were observed by the light scattering of a He–Ne laser beam by the desorbed particles. Recently, LIAD was successfully applied to desorb cells, viruses, microparticles, and nanoparticles with good efficiencies.^{56,67,68,71,72} To prepare the sample for LIAD, an aliquot (10 μL) of the purified particle suspension (containing 1 × 10⁷ particles per mL) was deposited onto a 375–425 μm thick Si wafer and dried in a desiccated box. The air-dried Si wafer was then positioned in the gap between the ring and end-cap electrodes of the QIT. A frequency-doubled pulsed Nd:YAG laser (Laser Technique, Berlin, Germany) with a wavelength of 532 nm and a laser energy of 30 mJ per pulse was used to irradiate the sample from the back side of the wafer. The laser duration was 6 ns, and the power density on the Si wafer ranged from 1 to 5 × 10⁸ W cm⁻². The strong absorption of laser photons induced an acoustic wave to cause particle desorption from the opposite side.⁵⁸ Desorbed charged particles were captured by the ion trap with an audio frequency power amplifier applied to the ring electrode. Helium served as the buffer gas in the trap center at 60 mTorr.

Fig. 5 A schematic of the experimental setup for laser-induced acoustic desorption mass spectrometry (LIAD-MS), which includes a quadrupole ion trap (QIT), a pulsed YAG laser, and a charge detector.

Fig. 6 Laser-induced acoustic desorption of aluminum oxide (Al₂O₃) particles. The figures show desorption at different laser energies.

Particle mass spectrometer

Instrumentation

The purpose of building the LIAD-MS was to measure the masses of cells/microparticles and virus/nanoparticles with high speed. For this device, the m/z ratios of the particles are measured by scanning the trapping frequency to eject charged particles. A Faraday disc is used to measure the number of charges (z) on each particle so that the mass of the charged particle can be determined. There are a few specific features for the mass spectrometer to be able to measure cells/microparticles compared to most commercial mass spectrometers. These features include the following: (1) LIAD is used to desorb charged particles into the quadrupole ion trap, (2) a very high m/z ratio can be obtained using a frequency scan with mass-selective instability mode, (3) a phase lock system is used to enhance the trapping efficiency,⁷³ (4) because laser firing is synchronized with a zero radio frequency (RF) voltage by a four channel digital delay/pulse generator, the direct measurement of the number of charges (z) on a particle without secondary electron amplification is possible, and (5) a corona discharge is used to enhance the number of charges on the particles. We also shielded the charge detector to reduce the electronic background.⁷²

Corona discharge

A pressure-controlled corona discharge is often important in this study because it enhanced the number of charges on a particle. Without the corona discharge, only tens to hundreds of charges can be obtained during LIAD for a single red blood cell.⁶⁸ For corona discharge, the He pressure was controlled to induce high-voltage discharge during the laser desorption process. Clear blue and white discharge plasmas were observed between the ion trap and desorption plate during the time in which the laser beam was incident on the back side of the Si wafer, as shown in Fig. 5(a) and (b). The typical pressure in the trap used in this experiment was approximately 50–150 mTorr to allow mild discharge without breakdown.

Mass analyzer – quadrupole ion trap

The quadrupole ion trap (QIT) was used for ion storage in which ions can be confined for a period of time and for ion selection with a frequency scan to cover high m/z ratios. The motion of the ions in the field of an ion trap can be described by the Mathieu equation. The following equation can be used to describe the relationship between the m/z ratio of the trapped ions with the RF frequency (ω) and the DC voltage (V).

Because the m/z ratio is linearly proportional to the DC voltage (V) and the DC voltage is typically confined to less than 10 000 V, the range of m/z ratios is limited to 3 orders of magnitude. For most commercial quadrupole mass spectrometers, the RF frequency is fixed at 1 MHz, and the maximum m/z is less than 6000. In contrast, m/z is inversely proportional to the square of RF frequency (ω). Because the RF frequency can be precisely controlled and varied from single digits to several megahertz, the m/z range can span more than 10 orders of magnitude when V is fixed and ω is scanned. With this approach, different sizes of particles with masses reaching 10¹⁷ Da can be trapped. A typical mass spectrum from the frequency scan to obtain the mass of each individual microparticle is shown in Fig. 7. The electronic noise is typically equivalent to ∼500 charges.

Fig. 7 The mass spectra of Jurkat cells. The time axis was converted to m/z, and the induced voltage on the y axis was converted to charge number.

Detection method – charge detector

The image charge collection time of a large particle is estimated to be on the scale of milliseconds, assuming the Faraday disc starts to “see” the charge when the particle is approximately 1 cm from it. The details of the charge detector are shown in Fig. 8. The detector board was designed to be mechanically compatible with the ion trap. A circular metal disc in the center of the detector board acted as a Faraday disc to collect the image charge induced by the particles.

Fig. 8 The circuit design of the charge detector: (a) electronic schematic and (b) a photograph of the detector with the components arranged on a 44 mm by 44 mm PCB board.

The Faraday disc is less than 1 cm from the exit hole of the ion trap when the detector is mounted under the ion trap. A stainless mesh with >90% transmission was placed between the Faraday disc and the exit hole of the ion trap to shield the AC field of the ion trap voltage. The diameter of the Faraday disc was 1 cm, which should be sufficient to cover all of the particles that emerged from the ion trap. The key parameters determined by the electronic calibration are listed in Table 1.

Table 1 The key parameters of a charge detector

Parameters	Value
Charge conversion ratio	∼50 electrons per mV
Noise voltage	∼10 mV rms
Equivalent noise electron	∼500 e rms
Integrator discharge time constant	∼10 ms

---

Mass resolution

The mass resolution is usually defined as R = m/Δm, where m represents the mass at the peak position and Δm represents the full width half maximum of the peak. For most small molecules, all of the molecules have the same mass if only one isotope exists in each element in the compound. Therefore, Δm represents the resolution of the instrument. However, for most bio-organic micro/nanoparticles, the mass distributions can be very broad or even larger than the instrumentation resolution. Under this circumstance, the mass spectrum can be used to measure the mass distribution of a particle. For charge detection by an ion trap mass spectrometer for the particle analysis described above, the resolution depends on the charge measurement and the ejection process. It is always useful to know the mass resolution of a mass spectrometer. Whitten et al.⁷⁴ and Goeringer et al.⁷⁵ reported the pressure effect on the mass resolution of an ion trap with resonance ejection. Up to now, few studies have investigated the resolution for microparticles measured under an axial mass-selective instability mode by scanning the trap-driving frequency. Because it is very difficult to get a microparticle sample with a sufficiently small size distribution, the mass resolution of an instrument is difficult to obtain. Nevertheless, by measuring a sample of polystyrene microparticles with a size of 3 μm and a size distribution less than 1% from NIST (SRM 1692), a mass resolution of 4 was obtained for masses of ∼10¹³ Da.⁷⁶

Mass accuracy

In term of the mass accuracy of the system, the measurement of NIST 3 μm polystyrene particles was repeated ten times. There were more than 1000 peaks obtained under the same experimental conditions. The standard deviation of the size of these 3 μm particles was 0.0028 μm. The accuracy of the instrument was approximately 2%. The data are shown in Table 2. Occasional “dimer” particles (a cluster with two microparticles) were trapped. Since the number of charges and the mass of a “dimer” is approximately twice that of a single particle, the m/z should be approximately the same as that of a single microparticle. Nevertheless, the amplitude corresponding to the total charges should be two-fold higher. The mass obtained can be determined as a dimer. A conventional mass spectrometer, other than some ion mobility spectrometers, cannot distinguish between M₂²⁺ and M⁺ because no charge information can be obtained and m/z is identical for both types of ions.

Table 2 The accuracy determination from the measurements of NIST 3 μm polystyrene by LIAD-MS

Repeat times	The mass result of NIST polystyrene 3 μm (Da)
1	7.27 × 10^12 ± 1.83 × 10^12
2	7.30 × 10^12 ± 1.82 × 10^12
3	7.34 × 10^12 ± 2.20 × 10^12
4	7.00 × 10^12 ± 2.09 × 10^12
5	7.13 × 10^12 ± 1.97 × 10^12
6	7.21 × 10^12 ± 1.99 × 10^12
7	7.34 × 10^12 ± 1.86 × 10^12
8	7.05 × 10^12 ± 1.97 × 10^12
9	7.34 × 10^12 ± 2.00 × 10^12
10	7.32 × 10^12 ± 1.99 ×10^12
Mass average	7.23 × 10^12 Da
S.D.	1.45 × 10^11 Da
Accuracy	∼2%

---

Sensitivity

For cell/microparticle measurements, the concentration was usually higher than 1 × 10⁶ particles per mL. An aliquot of the particle suspension was deposited onto a sample plate. The average density of particles on the sample plate was approximately 200 particles per mm.² Approximately 50 particles were desorbed into the ion trap in a single laser shot, and the mass spectrum was obtained after the accumulation of five laser shots. For the measurements of viruses/nanoparticles, the concentration used was usually higher than 1 × 10⁹ particles per mL; therefore, the average density of particles on the sample plate was ∼200 000 particles per mm². There were approximately 50 000 particles desorbed into the ion trap in a single laser shot, and the mass spectrum was often obtained after the accumulation of ten laser shots.

Standard microparticle and mass distribution measurements

The mass and charge distributions for polystyrene particles with sizes of 3, 7, 10, 15, and 30 μm are shown in Fig. 9. Standard polystyrene microparticles with monodisperse 7 and 10 μm are commercially available from Sigma-Aldrich Inc. The 3 and 30 μm particles can be purchased as standard reference material from NIST. The average masses were measured as 9.6 × 10¹² (3 μm), 1.4 × 10¹⁴ (7 μm), 3.3 × 10¹⁴ (10 μm), 8.3 × 10¹⁴ (15 μm), and 6.8 × 10¹⁵ Da (30 μm), which agree with the calculated masses of 8.8 × 10¹², 1.2 × 10¹⁴, 3.4 × 10¹⁴, 1.1 × 10¹⁵, and 8.6 × 10¹⁵ Da, respectively. The number of charges increases as the size increases but is not proportional to the surface area of a microparticle.

Fig. 9 The mass and charge distributions for various sizes of polystyrene microparticles. Each count represents a single detected microparticle. Fewer counts were obtained for the 30 μm microparticles (e and j) because it is more difficult to trap large particles due to gravity.

Mass measurement of cells

In addition to the measurement of the polystyrene microparticles, the mass distributions of various types of cells were measured. T lymphocytes (CD3⁺ cells) and monocytes (CD14⁺ cells) are major components of peripheral blood mononuclear cells, which play a critical role in the immune system. LIAD-MS was used to measure the mass distributions of lymphocytes and monocytes. The peak positions were determined to be 2 × 10¹³ and 4.2 × 10¹³ Da for the lymphocytes and monocytes, respectively (Fig. 10). A high number of charges (100 000 electrons) attached to cells in the ion trap was observed. The histograms of the masses and charges of the particles are shown in Fig. 11.

Fig. 10 The mass distributions of lymphocytes (CD3⁺ cells, black) and monocytes (CD14⁺ cells, red).

Fig. 11 The mass and charge distributions for different types of cells. Each count represents a single detected cell.

The relationship between the measured mass and charges of the polystyrene spheres and cells is shown in Fig. 12. The number of charges on a microparticle has a nearly linear relationship with the mass for particle masses less than 1 × 10¹⁵ Da.

Fig. 12 The averaged charge distribution vs. mass of five polystyrene spheres (black squares), two normal T-cells (green triangles), and two cancer cells (red circles).

Nanoparticle uptake by mammalian cells

To verify that the LIAD-MS could quantify the cellular uptake of nano/microparticles, NTERA2 cells were incubated with gold nanoparticles (NPs) with diameters of 30 and 250 nm, and their mass spectra were obtained. The cellular uptake of the 30 and 250 nm gold NPs as a function of incubation time is shown in Fig. 13. The difference in the mean mass was measured to be approximately 60% after incubating the NTERA2 cells with the 30 nm gold NPs for 24 hours (Fig. 13a). Fig. 13b shows a mean-mass deviation of 200%, which corresponds to an uptake of approximately ∼800 of the 250 nm gold NPs by the NTERA2 cells after incubation for 24 hours.

Fig. 13 The kinetics of the cellular uptake of (a) 30 nm and (b) 250 nm gold nanoparticles.

The kinetics of the uptake of the 50 nm gold NPs into NTERA2 cells as measured by CMS and ICP-MS (Fig. 14) were also studied. Both inductively coupled plasma mass spectrometer and charge-monitoring mass spectrometers identified a similar uptake trend and amount. Similar to the quantitative measurement of the uptake of gold NPs, CMS is expected to be a valuable tool for the quantification of all types of nanoparticles,⁷⁷ including polymeric NPs,⁷⁸ liposomes, viral-based NPs, carbon nanotubes, diamond NPs,⁷⁹ and polymeric micelles.

Fig. 14 (a) The extent of the cellular uptake of 50 nm gold nanoparticles by NTERA2 cells as a function of the incubation time as determined by a charge-monitoring mass spectrometer (red circles) and ICP-MS (black squares). (b) The extent of the cellular uptake of 100 nm polystyrene particles by Raw264.7 cells as a function of the incubation time. (The inserted photos are TEM images.)

Virus/nanoparticle detection

Instrumentation setup

The scheme for the experiment is shown in Fig. 15. This instrument is a modified version of the one shown in Fig. 6.^67,76 The structure is briefly described with an emphasis on the differences from Fig. 6. LIAD is used to desorb nanoparticles/viruses. A quadrupole ion trap is used to trap the desorbed charged particles. A phase lock system to enhance the trapping efficiency and to collect more desorbed particles was installed.⁷³ The laser firing is synchronized with the zero radio frequency (RF) voltage using a digital delay/pulse generator.

Fig. 15 A block diagram of the experimental setup, including a quadrupole ion trap, a pulsed Nd:YAG laser, a charge detector, a stainless steel shielding case, and a SiO₂ sample plate (400 nm thickness, high-resistance surface). An aliquot (10 μL) of the purified particles was placed on the front side of the sample plate. A frequency-doubled Nd:YAG laser beam (λ = 532 nm, 30 mJ per pulse) with a pulse duration of approximately 6 ns is directly incident on the back side of the sample plate. A synchronization of laser firing with the phase of the RF is installed to optimize the trapping efficiency.

Standard nanoparticles and mass distribution measurement

To check the feasibility of this instrument for measuring the masses of nanoparticles/viruses, standard polystyrene spheres with sizes of 50 and 100 nm were first measured. The calculated masses were 4.14 × 10⁷ and 3.3 × 10⁸ Da for the 50 nm and 100 nm polystyrene particles, respectively. The identification of these peaks was based on an approach similar to ESI with the assumption that there was an integer number of charges with the monomer and clusters of the monomer. These results are in good agreement with the 50 nm and 100 nm polystyrene spherical monoparticles with a mass of 4.14 × 10⁷ Da (Fig. 16).

Fig. 16 The mass spectra of (a) 50 nm and (b) 100 nm standard polystyrene spherical particles by MS.

Human Immunodeficiency Virus (HIV)

The mass of HIV-based lentivirus was also measured. The HIV-based lentivirus is a spherical virus approximately 90–120 nm in diameter with a three-layer structure and a virion buoyant density of 1.16–1.18 g cm⁻³ in sucrose.⁸⁰ The m/z of the HIV lentivirus was measured to be 3.51 × 10⁸ (M⁺, monomer), as shown in Fig. 17. These results are in good agreement with the calculated mass range of a single HIV lentivirus particle. The mass distribution of the HIV virus was approximately 10% (ΔM/M, monomer). Therefore, the observed mass variety should exhibit the mass distribution of the virus particles. Thus, rapid and reliable measurement of both the mass of a nanoparticle/virus and its mass distribution was achieved.

Fig. 17 The mass spectrum of HIV. The typical trapping parameters for Ω/2π and V_p–p are 1500 Hz and 1000 V, respectively. The scan time was 100 ms.

Perspective

Mass spectrometry now has the capability to detect ions with masses from the atomic to cellular ranges. However, the mass resolution for high mass particles is a concern. After several decades of effort, ultrahigh resolution FT-ICR was successfully used to demonstrate a resolution of 24 000 000 for protonated reserpine (m/z 609.28066).^81,82 Nevertheless, the mass resolution usually decreases very fast when the m/z is higher than 10 000. Therefore, additional work is needed to obtain sufficient resolution for high m/z ratios. High-resolution mass spectrometry in the high molecular weight region can assist in the identification of large biomolecules, especially top-down proteomic analysis. The post-translation modification (PTM) of protein makes the need for higher resolution more critical to distinguish different post-translation modifications for the same protein molecule. Assuming each protein has ten glycosylation sites and each glycosylation has five glycan forms, there can be more than 50 different glycoproteins for a single specific protein. If various isotopes, such as ¹³C and ¹⁵N, are taken into account, the mass resolution must be very high to separate different molecules. Otherwise, the spectrum will be very broad. Therefore, the mass resolution in the high-mass regime for mass spectrometry must be improved significantly. The new quadrupole ion trap mass spectrometer has a resolution power of 100–300 for a protein with a mass of 66.4 kDa.⁸³ Using low q_z resonance excitation ejection with a high trapping frequency (1.1 MHz), the resolution was improved. The high frequency and high voltage of RF create a sharp and deeper potential well for the ion trap. A modified Q-TOF instrument was used to improve the mass-resolving power in the high molecular weight region.^84,85 A linear ion trap was used to confine the initial kinetic energy spread when the ions were ejected out of the end-cap of the ion trap. The ejected ions were lined up with the minimum initial space and energy spread for an orthogonal time-of-flight mass spectrometer. A pulsed voltage then repelled the ions toward the detector. A new modified Orbitrap was also used to measure the large protein in the Q-TOF experiment.⁸⁶ The modified Orbitrap provides a significant improvement in the resolution. Single ion detection was performed for each single scan in the Orbitrap experiment. The overall resolution of the spectrum was obtained from the sum of 5000 single scans. The Orbitrap has been used to achieve a resolution of 5000 at an m/z of 25 000 and a resolution of 16 000 at an m/z of 10 000. Nevertheless, there have not been any reports on obtaining resolutions higher than 100 for particles with m/z ratios higher than 1 000 000. Microparticles can be detected by CMS, but the resolution is only ∼4. Designing a device that improves the resolution in the ultrahigh mass regime would be valuable for future studies.

Another concern is the absolute quantization of analytes. Matrix effects influence the accuracy of quantitation by MALDI. Ion suppression is notorious for the reproducibility of MALDI.⁸⁷ For ESI-based quantitation, several technologies have been developed to improve quantitation, such as label-free, stable isotopic labeling, chemical labeling, isobaric mass tags and spiked internal standards. These technologies give the relative quantitation of protein in proteomic analysis. New strategies have been developed for ICP-MS protein quantitation. Absolute quantitation of metal in metal-bound proteins can allow quantitation of the protein. The specific quantitation of membrane proteins was available via antibody-tagged polymer-chelating lanthanide metals using a CyTOF mass spectrometer.⁸⁸ Although this method is an indirect measurement, it greatly improves the absolute quantification of protein by mass spectrometry.

Conclusion

We have reported several mass measurement methods for ultra-heavy organic particles. QIT coupled with MIA is available to determine the masses ranging from 100 KDa to 30 MDa. Masses ranging from 10⁶ to 10¹⁶ Da for nano- and microparticles must be measured with a frequency-scanned QIT system. Mass spectrometry can now provide sensitive, rapid and routine analysis of large particles. Mass spectrometry has a demonstrated analysis capability from the atomic to cellular level, and it is a powerful tool in life science research. Challenges remain in the quantification and resolution in the ultrahigh mass region. A high-resolution instrument can provide more detailed structural information. Direct quantitation using the intensity of the mass spectra without internal standards will be extremely valuable for biological sample analysis, such as top-down PTM-proteomics.

References

1. F. W. Aston, Philosophical Magazine Series 6, 1919, 38, 707–714.
2. W. Bleakney, Phys. Rev., 1930, 35, 1180–1186.
3. J. H. Beynon, D. O. Jones and R. G. Cooks, Anal. Chem., 1975, 47, 1734–1738.
4. M. S. B. Munson and F. H. Field, J. Am. Chem. Soc., 1966, 88, 2621–2630.
5. H. D. Beckey, E. Hilt and H. R. Schulten, J. Phys. E: Sci. Instrum., 1973, 6, 1043.
6. R. Macfarlane and D. Torgerson, Science, 1976, 191, 920–925.
7. M. Barber, R. S. Bordoli, R. D. Sedgwick and A. N. Tyler, Nature, 1981, 293, 270–275.
8. J. S. Allen, Phys. Rev., 1939, 55, 966–971.
9. J. L. Wiza, Nucl. Instrum. Methods, 1979, 162, 587–601.
10. J. Griffiths, Anal. Chem., 2008, 80, 5678–5683.
11. S. F. Wong, C. K. Meng and J. B. Fenn, J. Phys. Chem., 1988, 92, 546–550.
12. M. Karas and F. Hillenkamp, Anal. Chem., 1988, 60, 2299–2301.
13. K. Tanaka, H. Waki, Y. Ido, S. Akita, Y. Yoshida, T. Yoshida and T. Matsuo, Rapid Commun. Mass Spectrom., 1988, 2, 151–153.
14. S. Jespersen, W. M. A. Niessen, U. R. Tjaden, J. van der Greef, E. Litborn, U. Lindberg, J. Roeraade and F. Hillenkamp, Rapid Commun. Mass Spectrom., 1994, 8, 581–584.
15. U. Boesl, J. Grotemeyer, K. Walter and E. W. Schlag, Instrum. Sci. Technol., 1987, 16, 151–171.
16. K. Ludwig, S. Habbach, J. Krieglstein, S. Klumpp and S. König, PLoS One, 2011, 6, e23612.
17. W. Shockley and J. R. Pierce, Proceedings of the Institute of Radio Engineers, 1938, 26, 321–332.
18. X. Tang, R. Beavis, W. Ens, F. Lafortune, B. Schueler and K. G. Standing, Int. J. Mass Spectrom. Ion Processes, 1988, 85, 43–67.
19. P. W. Geno and R. D. Macfarlane, Int. J. Mass Spectrom. Ion Processes, 1989, 92, 195–210.
20. G. Westmacott, W. Ens and K. G. Standing, Nucl. Instrum. Methods Phys. Res., Sect. B, 1996, 108, 282–289.
21. G. Westmacott, M. Frank, S. E. Labov and W. H. Benner, Rapid Commun. Mass Spectrom., 2000, 14, 1854–1861.
22. G. W. Fraser, Int. J. Mass Spectrom., 2002, 215, 13–30.
23. C. Bich, M. Scott, A. Panagiotidis, R. J. Wenzel, A. Nazabal and R. Zenobi, Anal. Biochem., 2008, 375, 35–45.
24. U. Riek, R. Scholz, P. Konarev, A. Rufer, M. Suter, A. Nazabal, P. Ringler, M. Chami, S. A. Müller, D. Neumann, M. Forstner, M. Hennig, R. Zenobi, A. Engel, D. Svergun, U. Schlattner and T. Wallimann, J. Biol. Chem., 2008, 283, 18331–18343.
25. C. A. Ross and M. A. Poirier, Nat. Med., 2004, 10, S10–S17.
26. C. H. W. Chen, L. J. Sammartano, N. R. Isola and S. L. Allman, AIP Conf. Proc., 2001, 584, 96–105.
27. C. Bich and R. Zenobi, Curr. Opin. Struct. Biol., 2009, 19, 632–639.
28. S. A. Trauger, T. Junker and G. Siuzdak, in Modern Mass Spectrometry, Springer, 2003, pp. 265–282.
29. T. Damian, Rep. Prog. Phys., 1996, 59, 349.
30. W. H. Benner, Anal. Chem., 1997, 69, 4162–4168.
31. A. A. Aksenov and M. E. Bier, J. Am. Soc. Mass Spectrom., 2008, 19, 219–230.
32. R. J. Wenzel, U. Matter, L. Schultheis and R. Zenobi, Anal. Chem., 2005, 77, 4329–4337.
33. E. B. Mark and J. G. Joseph, Active Learning Using the Virtual Mass Spectrometry Laboratory, American Chemical Society, 2007.
34. A. O. D. M. Sipe, B. Firek, R. Hendrix and M. E. Bier, ASMS2008, Proteomics: Applications, 2008, p. 571.
35. U. Bahr, U. Röhling, C. Lautz, K. Strupat, M. Schürenberg and F. Hillenkamp, Int. J. Mass Spectrom. Ion Processes, 1996, 153, 9–21.
36. D. C. Imrie, J. M. Pentney and J. S. Cottrell, Rapid Commun. Mass Spectrom., 1995, 9, 1293–1296.
37. D. Twerenbold, D. Gerber, D. Gritti, Y. Gonin, A. Netuschill, F. Rossel, D. Schenker and J.-L. Vuilleumier, Proteomics, 2001, 1, 66–69.
38. I. Giaever, H. R. Hart, Jr and K. Megerle, Phys. Rev., 1962, 126, 941–948.
39. E. W. Schlag, J. Grotemeyer and R. D. Levine, Chem. Phys. Lett., 1992, 190, 521–527.
40. C. M. Somers, B. E. McCarry, F. Malek and J. S. Quinn, Science, 2004, 304, 1008–1010.
41. T. M. Allen and P. R. Cullis, Adv. Drug Delivery Rev., 2013, 65, 36–48.
42. A. Gupta, D. Akin and R. Bashir, Appl. Phys. Lett., 2004, 84, 1976–1978.
43. B. Ilic, Y. Yang and H. G. Craighead, Appl. Phys. Lett., 2004, 85, 2604–2606.
44. M. A. Cooper, F. N. Dultsev, T. Minson, V. P. Ostanin, C. Abell and D. Klenerman, Nat. Biotechnol., 2001, 19, 833–837.
45. A. J.-C. Eun, L. Huang, F.-T. Chew, S. F.-Y. Li and S.-M. Wong, J. Virol. Methods, 2002, 99, 71–79.
46. D. Li, J. Wang, R. Wang, Y. Li, D. Abi-Ghanem, L. Berghman, B. Hargis and H. Lu, Biosens. Bioelectron., 2011, 26, 4146–4154.
47. R. Wang and Y. Li, Biosens. Bioelectron., 2013, 42, 148–155.
48. C. J. Hogan, E. M. Kettleson, B. Ramaswami, D.-R. Chen and P. Biswas, Anal. Chem., 2006, 78, 844–852.
49. F. Patolsky, G. Zheng, O. Hayden, M. Lakadamyali, X. Zhuang and C. M. Lieber, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 14017–14022.
50. C. S. Kaddis, S. H. Lomeli, S. Yin, B. Berhane, M. I. Apostol, V. A. Kickhoefer, L. H. Rome and J. A. Loo, J. Am. Soc. Mass Spectrom., 2007, 18, 1206–1216.
51. B. T. Ruotolo, J. L. Benesch, A. M. Sandercock, S.-J. Hyung and C. V. Robinson, Nat. Protoc., 2008, 3, 1139–1152.
52. S. D. Fuerstenau and W. H. Benner, Rapid Commun. Mass Spectrom., 1995, 9, 1528–1538.
53. J. C. Schultz, C. A. Hack and W. H. Benner, J. Am. Soc. Mass Spectrom., 1998, 9, 305–313.
54. J. C. Schultz, C. A. Hack and W. H. Benner, Rapid Commun. Mass Spectrom., 1999, 13, 15–20.
55. W. H. B. Stephen, D. Fuerstenau, J. J. Thomas, C. Brugidou, B. Bothner and G. Siuzdak, Angew. Chem., Int. Ed., 2001, 40, 541–544.
56. Z. Nie, Y.-K. Tzeng, H.-C. Chang, C.-C. Chiu, C.-Y. Chang, C.-M. Chang and M.-H. Tao, Angew. Chem., Int. Ed., 2006, 45, 8131–8134.
57. B. Lindner, Int. J. Mass Spectrom. Ion Processes, 1991, 103, 203–218.
58. V. V. Golovlev, S. L. Allman, W. R. Garrett and C. H. Chen, Appl. Phys. Lett., 1997, 71, 852–854.
59. V. V. Golovlev, S. L. Allman, W. R. Garrett, N. I. Taranenko and C. H. Chen, Int. J. Mass Spectrom. Ion Processes, 1997, 169–170, 69–78.
60. J. Pérez, L. E. Ramírez-Arizmendi, C. J. Petzold, L. P. Guler, E. D. Nelson and H. I. Kenttämaa, Int. J. Mass Spectrom., 2000, 198, 173–188.
61. R. C. Shea, C. J. Petzold, J. L. Campbell, S. Li, D. J. Aaserud and H. I. Kenttämaa, Anal. Chem., 2006, 78, 6133–6139.
62. R. C. Shea, S. C. Habicht, W. E. Vaughn and H. I. Kenttämaa, Anal. Chem., 2007, 79, 2688–2694.
63. M. A. Philip, F. Gelbard and S. Arnold, J. Colloid Interface Sci., 1983, 91, 507–515.
64. S. Schlemmer, J. Illemann, S. Wellert and D. Gerlich, J. Appl. Phys., 2001, 90, 5410–5418.
65. W.-P. Peng, Y.-C. Yang, M.-W. Kang, Y. T. Lee and H.-C. Chang, J. Am. Chem. Soc., 2004, 126, 11766–11767.
66. S. Schlemmer, S. Wellert, F. Windisch, M. Grimm, S. Barth and D. Gerlich, Appl. Phys. A, 2004, 78, 629–636.
67. W.-P. Peng, H.-C. Lin, H.-H. Lin, M. Chu, A. L. Yu, H.-C. Chang and C.-H. Chen, Angew. Chem., Int. Ed., 2007, 46, 3865–3869.
68. W.-P. Peng, Y.-C. Yang, M.-W. Kang, Y.-K. Tzeng, Z. Nie, H.-C. Chang, W. Chang and C.-H. Chen, Angew. Chem., Int. Ed., 2006, 45, 1423–1426.
69. Y.-F. Hsu, J.-L. Lin, S.-H. Lai, M.-L. Chu, Y.-S. Wang and C.-H. Chen, Anal. Chem., 2012, 84, 5765–5769.
70. J. L. Campbell, M. N. Fiddler, K. E. Crawford, P. P. Gqamana and H. I. Kenttämaa, Anal. Chem., 2005, 77, 4020–4026.
71. W.-P. Peng, Y.-C. Yang, C.-W. Lin and H.-C. Chang, Anal. Chem., 2005, 77, 7084–7089.
72. H.-C. Lin, J.-L. Lin, H.-H. Lin, S.-W. Tsai, A. L. Yu, R. L. C. Chen and C.-H. Chen, Anal. Chem., 2012, 84, 4965–4969.
73. C.-H. Chen, J.-L. Lin, M.-L. Chu and C.-H. Chen, Anal. Chem., 2010, 82, 10125–10128.
74. W. B. Whitten, P. T. A. Reilly and J. M. Ramsey, Rapid Commun. Mass Spectrom., 2004, 18, 1749–1752.
75. D. E. Goeringer, W. B. Whitten, J. M. Ramsey, S. A. McLuckey and G. L. Glish, Anal. Chem., 1992, 64, 1434–1439.
76. W.-P. Peng, H.-C. Lin, M.-L. Chu, H.-C. Chang, H.-H. Lin, A. L. Yu and C.-H. Chen, Anal. Chem., 2008, 80, 2524–2530.
77. K. Cho, X. Wang, S. Nie, Z. Chen and D. M. Shin, Clin. Cancer Res., 2008, 14, 1310–1316.
78. H. J. Johnston, M. Semmler-Behnke, D. M. Brown, W. Kreyling, L. Tran and V. Stone, Toxicol. Appl. Pharmacol., 2010, 242, 66–78.
79. B. Zhang, Y. Li, C.-Y. Fang, C.-C. Chang, C.-S. Chen, Y.-Y. Chen and H.-C. Chang, Small, 2009, 5, 2716–2721.
80. C. M. Fauquet, M. A. Mayo, J. Maniloff, U. Desselberger and L. A. Ball, Virus Taxonomy: Classification and Nomenclature of Viruses, Elsevier Academic Press, San Diego, California, USA, 2005.
81. G. Squires, J. Chem. Soc., Dalton Trans., 1998, 0, 3893–3900.
82. E. Nikolaev, I. Boldin, R. Jertz and G. Baykut, J. Am. Soc. Mass Spectrom., 2011, 22, 1125–1133.
83. G. E. Reid, J. M. Wells, E. R. Badman and S. A. McLuckey, Int. J. Mass Spectrom., 2003, 222, 243–258.
84. R. H. H. van den Heuvel, E. van Duijn, H. Mazon, S. A. Synowsky, K. Lorenzen, C. Versluis, S. J. J. Brouns, D. Langridge, J. van der Oost, J. Hoyes and A. J. R. Heck, Anal. Chem., 2006, 78, 7473–7483.
85. J. Lee, H. Chen, T. Liu, C. E. Berkman and P. T. A. Reilly, Anal. Chem., 2011, 83, 9406–9412.
86. R. J. Rose, E. Damoc, E. Denisov, A. Makarov and A. J. R. Heck, Nat. Methods, 2012, 9, 1084–1086.
87. R. Knochenmuss, F. Dubois, M. J. Dale and R. Zenobi, Rapid Commun. Mass Spectrom., 1996, 10, 871–877.
88. D. R. Bandura, V. I. Baranov, O. I. Ornatsky, A. Antonov, R. Kinach, X. Lou, S. Pavlov, S. Vorobiev, J. E. Dick and S. D. Tanner, Anal. Chem., 2009, 81, 6813–6822.

---