Introduction

Biosensors and clinical diagnostic tools based on microfluidics are now commercially available for point-of-care diagnosis [1, 2]. These devices rely usually on three major detection methods: spectroscopic, electrochemical, and mass spectrometry (MS) methods [3]. Despite their simplicity and high sensitivity, optical and electrochemical methods often require labeling procedures using tags with optical or electrochemical properties. The indirect sensing process can cause false-positive signals and cannot provide exact information about the analyte itself [4, 5]. In contrast, MS provides direct, highly specific, and accurate characterization and quantification of target analytes and thereby has gained increasing importance as a detection method for microfluidic chips [6]. However, compared to optical and electrochemical methods, which can be integrated in chips, a key difficulty in coupling microfluidics with MS is the design of the emitter [7].

Compatible with the small volumes of samples used in microfluidic chips, nanoelectrospray ionization (nanoESI) is one of the most suitable interfaces for integrating microfluidics with MS, as it provides good detection sensitivity at submicroliter per minute flow rates [8]. NanoESI interfaces between MS and microfluidic chips can be classified in three types: (i) spraying from the microchannel outlet [9, 10], (ii) spraying from an externally coupled emitter [11], (iii) spraying from an emitter integrated on chip [12, 13]. Electrospray directly from a microfluidic channel is the most straightforward interfacing approach used in early days, but the performance is limited by the solution spreading at the nontapered and hydrophilic flat edge of the channel [9, 10]. Many approaches have been followed to address this problem. For example, Bedair et al. constructed a polymeric monolith at the edge of an open channel to spray samples from the pores [14]. Another option is to integrate the microfluidic chip with an external nanoESI emitter, which is normally obtained from a conventional pulled glass capillary with a diameter of several micrometers to several hundred nanometers [15, 16]. However, the dead volume originating from the external connection is inevitable. Clogging of the emitter tip is also a disturbing aspect for stable performance [17]. As a result, on-chip integrated emitters have become popular and a wide range of fabrication methods have been employed to build them [18,19,20,21]. Microchannel created by laser ablation as an ESI emitter is perhaps the easiest approach [13], where channel outlet was cut into a sharp-tip shape to assist stable electrospray ionization [22, 23]. Lower sample consumption and higher detection sensitivity could be obtained by decreasing the size of sample infusion channel or delivering sample solution into another dedicated nanoESI emitter channel. However, the continuous sample infusion will be impeded in both cases due to the large backpressure caused by the small emitter channel with a length of several millimeters. An on-chip microhole as an emitter for electrostatic spray ionization was created for coupling MS with droplet-based microfluidics previously (spyhole-ESTASI). However, it was difficult to align the contactless electrode with the spyhole and MS inlet, and the spray performance was limited by the electrical field spreading due to the much larger size of the electrode compared with the spyhole.

To alleviate these problems, we present here a novel interface (spyhole-nanoESI) to couple microfluidic chips with MS by drilling a microhole with a diameter of 10–12 μm directly on top of a microchannel. Different from spyhole-ESTASI, an internal electrode contacting the sample channel was integrated on the chip to induce nanoESI for higher sensitivity and better spray performance. Under optimal conditions, the flow rate could be lowered to 100 nL/min for nanoESI. The dead volume of the microchip emitter is about 100 nL, and an easy way to reduce the sample consumption is to use an eight-way valve to split the flow and pressure. Limits of detections (LODs) could be obtained at attomole per microliter levels for small organic molecules and peptides, which is 600 times lower than that usually obtained with standard ESI and microESI. To take advantages of the high-sensitivity and low-dead volume, the spyhole-nanoESI was used to detect a small cell lung cancer (SCLC) biomarker protein of neuron-specific enolase (NSE). By monitoring the transition of the unique peptide of NSE, LOD for the biomarker by the spyhole-nanoESI was at 12 pM. Its potential application in cancer diagnosis was further demonstrated with 1 μL of human serum containing 2 nM NSE, which is in accordance with the real abundance of the biomarker in blood at the extensive stage of the lung cancer [24], without any complicated sample preparation steps, such as immunoaffinity extraction and liquid chromatography separation. In addition, the microchips can be disposable to avoid any cross contamination.

Experimental

Materials and Methods

Angiotensin I (Ang I, trifluoroacetate salt, 98%) was obtained from Bachem (Dübendorf, Switzerland). Cytochrome c (from horse heart), acetic acid (99.5%), trifluoroacetic acid (TFA), 1,4-dithiothreitol (DTT, ≥ 99%), and ammonium bicarbonate (NH4HCO3, 99.5%) were purchased from Fluka (Buchs, Switzerland). Methanol (99.9%), isopropanol (≥ 99.5%), reserpine (≥ 99%), iodoacetamide (IAA, 97%), insulin (from bovine pancreas, HPLC), and recombinant human neuron-specific enolase (NSE, ≥ 95%) were obtained from Sigma-Aldrich (Buchs, Switzerland). Trypsin (from bovine pancreas) was obtained from Applichem (Darmstadt, Germany). Human plasma was purchased from Bioreclamation LLC (Westbury, NY, USA). Deionized water produced by an alpha Q-Millipore System (Zug, Switzerland) was used for all experiments. A commercial silica emitter for nanoESI with a tip diameter of 15 μm was obtained from New Objective Inc. (Woburn, MA, USA). Pierce Top 12 abundant protein depletion spin columns were purchased from Thermo Scientific (Switzerland). Amicon Ultra-2 centrifugal filters were ordered from Merck Millipore (Schaffhausen, Switzerland). Aquapel® solution was obtained from Aquapel Glass Treatment (Pittsburgh, USA).

Spyhole-nanoESI MS/MS Experiment

A spyhole-nanoESI device was designed as shown in Fig. 1 and fabricated from a polyimide (PI) substrate (125 μm thick, Dupont, Switzerland) by scanning laser ablation as previously reported [25]. Briefly, a deeper short channel called reservoir was created with a depth of ~ 115 μm at the end of the sample channel. On top of this reservoir, a μ-hole with a diameter of 10–12 μm was drilled as the nanoESI emitter. More details about the fabrication process and the hydrophobic modification could be found in the Supporting Information (SI) section S1-1.

Figure 1
figure 1

Schematic representation of the spyhole-nanoESI device: a top-view, b side-view, and c 3D view. HV high voltage, LIT-MS linear ion trap mass spectrometry

Full size image

For MS analysis, the spyhole-nanoESI was mounted on a holder fixed on an x, y, and z positioning stage (Thorlabs, Dachau/Munich, Germany) and placed horizontally around 1 mm below a “L-shaped” ion transfer capillary, Fig. 1c and Fig. SI-1, instead of the original one of Thermo LTQ Velos (Thermo Scientific, San Jose, USA). Sample was infused at 100 nL/min with a syringe driven by a syringe pump. More details about the parameters for MS and tandem MS with collision-induced dissociation (CID) were given in SI-1.

For sensitivity comparison, four other setups, chiptip-ESI [22], spyhole-nanoESTASI [25], commercial Picotip®-nanoESI, and standard ESI, were also used. The design and fabrication of chiptip-ESI and spyhole-nanoESTASI were described in SI-3. The parameters for the MS analyses with these setups were also given in detail in SI-3. The applied voltage and flow rate were tailored accordingly for each setup to obtain an optimal signal.

NSE Tryptic Digestion, In Silico Digestion, and Unique Peptide Determination

Enzymatic digestion of NSE was performed with a commonly used procedure, including reduction, alkylation, and tryptic digestion. First, 30 μL NSE at 10 μg/mL and 100 ng/mL in 10 mM NH4HCO3 buffer (pH 8) was denatured at 95 °C for 5 min and then incubated with 0.5 μL of freshly prepared 60 mM DTT in H2O at 57 °C for 30 min, respectively. After being cooled to room temperature (RT), 0.5 μL of freshly prepared 30 mM IAA in H2O was added into the sample solutions and placed in dark for another 30 min at RT. For tryptic digestion, 0.5 μL trypsin solutions at 12 and 0.6 μg/mL in 10 mM NH4HCO3 buffer (pH 8) were added into the two samples containing 10 μg/mL and 100 ng/mL NSE, respectively, and incubated overnight at 37 °C with moderate shaking. The digestions were quenched by adding 0.5 μL TFA and the diluted digests in ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid) were ready to be directly analyzed by spyhole-ESI MS/MS without any additional procedure.

The sequence of NSE was acquired from the UniprotKB database with the accession number of P09104. MS-digest program from protein prospector version 5.19.1 (University of California, San Francisco, CA, USA) was used to perform an in silico digest of NSE. The results allowed the selection of possible unique peptide candidates by matching with the experimental data. To investigate if the candidate peptides solely originate from NSE, peptide search with Uniprot was performed to find all proteins in UniProtKB that contain the query peptide sequence. The search was limited to Homo sapiens and Bos taurus because trypsin was derived from bovine pancreas.

Detection of NSE in Human Serum

Serum was derived from human plasma by overnight clotting and centrifugation. Certain amounts of NSE were fortified into 30 μL serum to reach a final concentration of 200, 100, and 50 ng/mL, respectively. Firstly, the depletion of high-abundant proteins in serum was performed to reduce the sample complexity. Thirty microliter of the serum sample was pooled into a Pierce Top 12 abundant protein depletion spin column and incubated for 1 h at RT. Centrifugation at 1000×g was then performed with the spin column for 2 min to collect the depleted serum sample in 500 μL 10 mM phosphate-buffered saline (PBS) buffer (0.15 M NaCl, 0.02% azide, pH 7.4). Before tryptic digestion, 500 μL of the solution collected from the 30 μL depleted serum sample was desalted by an Amicon Ultra-2 centrifugal filter device with the nominal molecular weight limit of 10 K and finally concentrated in 30 μL 10 mM NH4HCO3 buffer (pH 8). After denaturation at 95 °C for 5 min, 1 μL of freshly prepared DTT (150 mM) was added and incubated at 57 °C for 30 min. Then, the sample solution was alkylated by 1 μL of freshly prepared IAA (75 mM) in dark for 30 min at RT, followed by digestion with 0.5 μL of 600 μg/mL trypsin (freshly prepared in 10 mM NH4HCO3) overnight at 37 °C with moderate shaking. After the digestion quenched by 0.5 μL TFA, the serum digests with or without fortified NSE were diluted ten times with ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid) and then analyzed directly with different methods, including the spyhole-nanoESI MS/MS, chiptip-ESI MS/MS, and standard ESI MS/MS without additional desalting, concentration or fractionation procedures. The MS parameters were described in SI-1 and SI-3.

Results and Discussion

Spyhole-nanoESI MS: Microchip Design and Setup Optimization

In previous work, a setup called spyhole-ESTASI was constructed for efficient coupling of droplet-based microfluidics with MS. A spyhole with a diameter of 50 μm was drilled by laser ablation on the top of a polyimide microchip. A disk electrode (Ø 1 mm) was placed below the chip under the spyhole and connected with a high-voltage generator to induce the electrostatic spray ionization (ESTASI) of water-in-oil droplets when passing through the microchannel for product monitoring. The working mechanism of ESTASI has been reported elsewhere, which could be considered as the charging and discharging processes of two-tandem capacitors, electrode/PI layer/sample, and sample/air/MS, respectively. To avoid the additional oil removal step, the microchip outlet remained open and the oil continued flowing to the outlet to be adsorbed by a swab stick during the MS analysis of droplets through the spyhole. Absence of direct contact of the electrode with the sample solution prevented any electrochemical reactions. However, it was difficult to align the contactless electrode with the spyhole and MS inlet, and the spray performance was limited by the electrical field spreading due to the much larger size of the electrode compared with the spyhole.

In the present technique of spyhole-nanoESI, a much smaller size of spyhole was created for higher sensitivity. An internal carbon electrode was integrated on the chip to induce nanoESI by contacting the sample solution in the microchannel to achieve better spray performance. For these purposes, two parameters of the spyhole, depth and width, were optimized. The spyhole depth was the distance between the spyhole outlet and the sample channel and should be as small as possible to facilitate the continuous sample delivery and avoid being susceptible to clogging. Instead of making the whole sample channel deeper, a short reservoir was created on the bottom of the sample channel to keep the total chip volume as small as possible to minimize dead volumes and sample consumption. Using a test sample of angiotensin I (Ang I) in ESI buffer (50% MeOH, 49% H2O and 1% acetic acid), the optimal spyhole depth was found to be 10 μm. A smaller depth of spyhole was not considered, because the short reservoir would then be easily penetrated during the laser ablation or the following treatment of the microchip. For nanoESI, the spyhole size plays a critical impact on the ionization efficiency and should be fabricated as small as possible. A spyhole with the diameter of 10–12 μm could be made reproducibly by laser ablation, as shown in Fig. 2a. In addition, the surface of the polyimide microchip was further modified to be hydrophobic by a commercial water-repellent Aquapel® solution, in order to form a stable and small droplet for nanoESI. Under optimal conditions, 0.15 nM of Ang I could be detected at 433 m/z with a signal-to-noise ratio (S/N) of 10 by the spyhole-nanoESI with an ion trap mass spectrometer (Thermo LTQ Velos), Fig. 2c.

Fig. 2
figure 2

Optimization of the spyhole width and depth of the spyhole-nanoESI microchip: a microscopic images of spyholes with diameters of 12, 25, and 35 μm. b Microscopic image of the charged droplet generated at the spyhole of 12 μm in diameter. c LODs of Ang I in ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid) by three spyhole-nanoESI microchips with different spyhole dimensions. The standard deviations of S/N values were obtained from three replicate measurements using the same chip. The MS detection conditions were optimized as given in SI-1

Full size image

The optimization of the spyhole-nanoESI MS with respect to hydrophobic modification, ESI voltage, the automatic gain control (AGC) target value of the ion trap instrument, and the vertical distance between the spyhole and the ion transfer capillary was given in SI-1 and SI-2. In summary, the optimized conditions were as follows: ESI buffer: 50% MeOH, 49% H2O, 1% Acetic acid; flow rate: 100 nL/min; ESI voltage: 3 kV, the AGC value: 10′000; vertical distance: 1 mm.

Sensitivity of Spyhole-nanoESI MS

Under the optimized conditions as described above, the sensitivity of spyhole-nanoESI was measured and compared to standard ESI from commercial ion source, PicoTip®-nanoESI from a commercial micro-emitter, and spyhole-nanoESTASI, which were described in detail in SI-3. For chiptip-ESI, the tip of the sample channel (110 μm × 50 μm) was cut into a V-shape as a spray emitter. Ionization with spyhole-nanoESTASI was achieved using a recently developed method of electrostatic spray ionization (ESTASI) [26]. The sample channel and the spyhole on the spyhole-nanoESTASI were the same as those of spyhole-nanoESI but without on-chip electrode for ESI. Instead, a disk metallic electrode (1 mm Ø) was placed below the microchip and in line with the spyhole and the “L-shaped” ion transfer capillary to induce ESTASI by applying a pulsed high voltage of 10 kV.

Ang I in ESI buffer was analyzed with spyhole-nanoESI, chiptip-ESI, spyhole-nanoESTASI, PicoTip®-nanoESI, and standard ESI, respectively. The LODs of Ang I based on the observation of its triply protonated ions (m/z = 433) were listed in Table 1. From the results, the spyhole-nanoESI demonstrated the best sensitivity with the S/N of 4 for 0.05 nM Ang I, as shown in Table 1 and Fig. 3b. In contrast, the LODs with chiptip-ESI, spyhole-nanoESTASI, commercial PicoTip®-nanoESI, and standard ESI were determined to be 5, 15, 0.2, and 30 nM, respectively, for Ang I with S/N ≥ 3. It is worth mentioning that different flow rates were used for different setups. For the spyhole-nanoESI, the upmost flow rate the spyhole allowed was smaller than 1.2 μL/min before the detachment of the lamination layer and a stable spray could still be obtained when lowering the flow rate to 100 nL/min. Lower flow rate than 100 nL/min was not further tested, which was limited by the current syringe pump. While for spyhole-nanoESTASI, the optimized flow rate was 200 nL/min; for chiptip-ESI, the optimized flow rate was 1 μL/min; for standard ESI, the optimized flow rate was 3 μL/min. Therefore, the spyhole-nanoESI showed LODs in terms of concentration better than the commercial PicoTip®-nanoESI, 600 times lower than standard ESI and 100 times lower than the previously developed chiptip-ESI [22]. The sample consumption is decreased by ~ 1000 times lower than chiptip-ESI and ~ 20,000 times lower than standard ESI by considering the difference in flow rate. The reproducibility and stability of spyhole-nanoESI were also evaluated using the same chip or various different chips. However, the signal with the S/N ≥ 3 at m/z 433 was hardly repeatable among various different spyhole-nanoESI chips for the detection of 0.05 nM Ang I. Instead, when the concentration of Ang I was increased to 0.15 nM, reproducible results were achieved using the same chip or various different chips, from the S/N values and standard deviations shown in SI-4. After 3 or 15 days’ storage at RT, the chips could still provide repeatable results for the analysis of 0.15 nM Ang I (data not shown), showing the good short-term and long-term stability of spyhole-nanoESI. However, this signal with the S/N ≥ 3 was hardly repeatable among different batch tests even with the same one chiptop-nanoESI for such an ultralow concentration of 0.05 nM. Instead, when the concentration of Ang I was increased to 0.15 nM, a reproducible signal with the S/N ≥ 5 could always be detected either with the same chip (three tests repeated) or different chiptop-nanoESI devices (> 4 duplicates). With such high performance, the spyhole-nanoESI can bring additional value to the coupling of microfluidics with MS. Considering that the ion injection time was around 1 ms to generate a mass spectrum for 0.05 nM Ang I, the infused Ang I at 100 nL/min was ~ 50 molecules to generate a mass spectrum, which was comparable with other reported nanospray devices for peptide detection [8, 27, 28].

Table 1 LODs of Ang I in ESI buffer (50% MeOH, 49% H2O, 1% acetic acid) with various devices. The detection conditions for the four different devices were described in SI-1 and SI-3. The standard deviations of S/N values were obtained from three replicate measurements using the same chip. I.D internal diameter
Full size table
Figure 3
figure 3

Detection of different molecules in ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid) by the spyhole-nanoESI MS: a 0.1 nM reserpine and b 0.05 nM Ang I. The detection conditions were the same as the optimized one given in SI-1

Full size image

The high sensitivity of the spyhole-nanoESI was also demonstrated with reserpine, a small drug molecule, with LOD as 0.1 nM, Fig. 3a. The good performance of the spyhole-nanoESI stems from the advantages of nanoESI [8], the well-controlled emitter geometry, and the surface hydrophobic modification of microfluidic chips, as discussed in detail in SI-5.

Detection of Cancer Biomarker with the Spyhole-nanoESI MS/MS

With the high sensitivity, the spyhole-nanoESI MS could be applied to detect cancer biomarkers directly from body fluids. As a proof-of-concept, a lung cancer biomarker protein, neuron-specific enolase (NSE), at low abundance, was detected by monitoring the transition of its unique peptide with the spyhole-nanoESI MS/MS.

Signature peptide candidates were selected from the peptides detected from the ten times diluted tryptic digests of 10 μg/mL NSE in ESI buffer. By matching the experimental data to the peptide list from in silico proteolysis of the protein and searching the sequences of the selected peptides with the UniprotKB proteome database, four unique peptides that could be easily detected by the spyhole-nanoESI MS were chosen as the signature peptide candidates. After optimizing the transition for each signature peptide candidate through single reaction monitoring (SRM) with the collision-induced dissociation (CID), a peptide IVIGMDVAASEFYR (namely SP-14 afterwards) was selected as the signature peptide specific for NSE, and the transition of 785.82+ → 843.4+ with the optimal normalized collision energy of 20 and the isolated width of 1 m/z was monitored for the following NSE detection. More details about the signature peptide selection were given in SI-6.

Under the conditions, the spyhole-nanoESI MS/MS was firstly used to detect NSE in water. With 10 ng/mL (0.2 nM) NSE tryptic digest, the fragment ion at m/z = 843.4 could be observed with good S/N of 50, as shown in Fig. 4a. In contrast, chiptip-ESI and standard ESI gave the S/N of 6 and 3 for 10 ng/mL NSE, respectively, as shown in Fig. 4b and c. As shown in Fig. 4d, the ESI buffer did not give any signal at m/z 834.4, confirming that the fragment ion solely originating from NSE instead of contaminates in the microchip or MS instrument.

Fig. 4
figure 4

NSE detection by single reaction monitoring (SRM) of its signature peptide SP-14. Ten nanogram per milliliter NSE tryptic digest in ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid) was used for the following tests. Mass spectra of the signature fragment ion of SP-14 at m/z 843.4 obtained by a spyhole-nanoESI MS/MS, b chiptip-ESI MS/MS, and c standard ESI MS/MS. d Mass spectrum of the ESI buffer obtained by spyhole-nanoESI MS/MS

Full size image

The LOD of NSE with the S/N of 3 by spyhole-nanoESI was then estimated as around 0.6 ng/mL (12 pM) and was comparable to that (0.4 ng/mL) reported with liquid chromatography (LC)-MS/MS based method [29]. In the literature, specific capture was firstly performed to collect the low abundant NSE (5–500 ng/mL, 1 mL), and then LC was applied to desalt and fractionate the tryptic peptides before MS/MS analysis. In contrast, the spyhole-nanoESI MS/MS requires no additional sample pretreatment prior to analysis and provides good sensitivity even without any pre-concentration step.

The spyhole-nanoESI MS/MS was further applied to detect the biomarker directly from human serum to demonstrate its potential application in clinical diagnosis. Human serum from healthy individuals was spiked with 0, 100, or 200 ng/mL of NSE for test. Then, the top 12 abundant proteins in the serum sample were depleted with immune spin columns and then digested by trypsin. The tryptic digests were directly analyzed with spyhole-nanoESI MS/MS after being diluted ten times with ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid). For the ten times diluted human serum without spiked NSE, there was no obvious signal of the signature fragment ion of SP-14 at m/z = 834.4 observed, as shown in Fig. 5c, while the ion could be found from the serum spiked with 200 ng/mL NSE (S/N = 10) or 100 ng/mL NSE (S/N = 3), Fig. 5a and b. Therefore, the elevated NSE at 100 ng/mL (2 nM) in 1 μL human serum could be detected with the spyhole-nanoESI MS/MS, which could be valuable for the fast and simple diagnosis of SCLC at the extensive stage, wherein the concentration of NSE could reach 94.5 ± 13.8 ng/mL [24], and could also be used as the early indicator of the response rate to chemotherapy. In contrast, with the chiptip-ESI and standard ESI, no obvious signal of the fragment ion of SP-14 could be detected either from normal human serum or the serum spiked with 100 ng/mL NSE.

Fig. 5
figure 5

Detection of NSE in human serum with spyhole-nanoESI MS/MS. One microliter of tryptic digests of human serum spiked with NSE was diluted by ten times with ESI buffer (50% MeOH, 49% H2O, and 1% acetic acid) for spyhole-nanoESI MS/MS analysis. Mass spectra of the signature fragment ion of SP-14 at m/z 843.4 from serum sample spiked with a 200 ng/mL NSE, b 100 ng/mL NSE, and c without NSE. The detection conditions were the same as those given in SI-2

Full size image

Compared with the routine immunometric method for NSE determination, the present method requires no usage of expensive labeled antibodies and provides exact identification, avoiding possible false-positive diagnosis from antibody cross reaction. Using LC-MS/MS-based methods for NSE detection, numerous cleanup processes before analysis are rather complicated, labor-extensive, time-consuming, and can cause sample loss. A large volume of serum from several hundred microliters to milliliters was usually required for specific separation and concentration. In contrast, the spyhole-nanoESI MS/MS could offer direct and sensitive diagnosis of the low abundant biomarker in human serum and requires only one cleanup procedure of high abundant protein depletion. Owing to the simplified procedure and great sensitivity, only 1 μL of serum was consumed for a single diagnosis.

In addition, the laser ablation method for microchip fabrication offers a precise process control, wherein the spyhole-nanoESI devices were reproducible. Its easy fabrication, low cost and disposable nature can eliminate cross contamination between samples, which is highly suitable for clinical application. Moreover, the spyhole-nanoESI device could be coupled with various MS instruments, such as high-resolution mass spectrometers for improved accuracy of analyte identification.

Conclusions

In the present work, a microchip device was coupled with MS by drilling a spyhole with a diameter of 10–12 μm on top of the chip to form a nanoESI emitter for ultrasensitive analysis. The sample was infused at a nanoscale flow rate and emitted from the hydrophobic-treated spyhole into MS by a nanoelectrospray ionization. With the optimized detection conditions, peptides were detected with ultrahigh sensitivities. The LOD of Ang I was as low as 50 pM, which was 600 times lower than standard ESI. The sample consumption for MS detection can be lowered by 20,000 times compared to standard ESI.

With the spyhole-nanoESI MS/MS, a lung cancer biomarker protein of NSE could be detected sensitively with LOD of 12 pM by monitoring the transition of its signature peptide. The biomarker can also be detected from human serum with concentrations as low as 2 nM. Only 1 μL of serum digests was sufficient for a single diagnosis. The great sensitivity, low cost, disposal nature, compatible with various MS instruments, together with the overall simplicity of the direct analysis with few cleanup processes make the spyhole-nanoESI MS/MS suitable for clinical diagnosis with a limited amount of samples.