This story originally ran on Aug. 3.
By Tony Fong
Name: Akshay Naik
Position: Research engineer, California Institute of Technology, 2008 to present
Background: Postdoc scholar, California Institute of Technology, 2006 to 2008; postdoc scholar, University of Maryland, 2006
Researchers at the California Institute of Technology have created a nanoscale mass spectrometer that they said can measure the mass of a single molecule in real time.
As research trends toward biological analyses at increasingly smaller scales, mass spectrometry with a sensitivity of a few molecules to a single one (seems grammatically more correct) will be necessary, according a study published June 21 in the online edition of Nature Nanotechnology describing the Cal Tech scientists' work.
In the study, the team of researchers demonstrate for the first time "mass spectrometry based on single biological molecule detection with a nanoelectromechanical system," or NEMS, mass spectrometer, they said.
In general, NEMS technology exploits the idea that the smaller a device, the more susceptible it will be to external disturbances. "This enhanced sensitivity of NEMS is opening a variety of unprecedented opportunities for applications such as mass spectrometry," the researchers said in their study.
"To reliably detect the expression of low-level signals and to understand the fundamental biological processes, it is important to develop techniques capable of single-cell or single-molecule analysis," they added.
In their paper, they utilized the mass sensitivity of ultra-high frequency NEMS resonators to demonstrate a "new paradigm" for mass spectrometry. The approach uses the resonators, which vibrate at a high frequency, to measure the mass of a molecule. When a molecule lands on the resonator, the frequency of the vibration changes, and based on that shift, a mass is measured, which can then be correlated with a protein.
According to the authors, the approach enables the first real-time detection of individual proteins and nanoparticles as they adsorb onto a sensitive NEMS detector.
Each NEMS sensor "in the single-molecule limit" acts essentially as a mass spec. "This NEMS-based mass spectrometry system, combined with other micro- and nanoscale technology, offers the possibility of compact, massively parallel MS, limited only by the number of NEMS mass sensors incorporated on a chip," they said.
Testing their technology on a sample of bovine serum albumin, which has a known mass of 66 kiloDaltons, they detected a frequency change of 1.2 kiloHertz, while beta-amylase protein, which has a mass of 200 kiloDaltons, caused a frequency shift of about 3.6 kHz.
ProteoMonitor recently spoke with Akshay Naik, the principal author of the study and a research engineer at Cal Tech, about his work developing the technology. Below is an edited version of the conversation.
Describe the concept of a nanomechanical system mass spectrometer and what it is you're trying to achieve.
The NEMS device that we used is a beam which is clamped at the two ends. The NEMS is about 2 microns long and is about 120 nanometers wide, vibrating at a resonant frequency, which is about 450 megaHertz.
When something lands on it — some protein or some molecule — its resonance frequency changes. We can use this principle to detect the mass of the protein.
What we did was electrospray the protein ion onto the NEMS and as the individual protein molecules landed on NEMS, the resonance frequency of the NEMS changed. And this frequency change is proportional to the mass of the individual protein molecule that's landing on the NEMS.
So using about 500 frequency shifts, we were able to extract some qualitative information about the composition of the proteins.
It's a first step toward doing a mass spectrometry using NEMS.
[ pagebreak ]
Are there specific challenges to developing this type of mass spectrometer for proteomics, and if yes, how did you approach this?
This is just a proof-of-principle experiment where we did the experiment with just one NEMS device, so in future generations … in five to 10 years, we are hoping we will have an array of devices using semiconductor fabrication techniques, and each of these NEMS devices would act as an individual mass spectrometer.
So you could imagine, let's say 10,000 NEMS devices and if each NEMS device could detect a single protein molecule in, let's say about 10 milliseconds, it could process the mass of about 10 million molecules in 100 seconds.
So that's a huge, huge advantage. But that would also be a challenge because you have to integrate the NEMS devices with measurements and you have to put all these things together.
Another challenge would be pushing the mass sensitivity, mass resolution levels down to a single Dalton … but we are working on it, and I think we should be able to get there in a few years.
Is the main advantage of the NEMS technology the sensitivity then?
The main advantage would be parallel measurement. Every single NEMS device, which is just a 2-micron long device, acts as a mass spectrometer, so you could have a … thousand, maybe even millions of these NEMS devices on a chip, and that makes it possible to do these highly parallel measurements.
All of these NEMS devices are acting as a mass spectrometer, and each of them would be doing a mass measurement, and they would be doing it for every individual molecule. They would be detecting every single molecule landing on it.
Another thing that's important is the huge mass range that this [device] has. In principle, it can detect from 1 Dalton all the way to 10 megaDaltons without changing anything. You don't have to change anything in the system, and you should be able to detect it in this entire range.
We haven't gone to a sensitivity level of a single Dalton … but from our sensitivity levels, which are about 10 kiloDaltons to 10 megaDaltons, it should be possible. Once we have reached 1 Dalton levels, I don't see any reason we can't detect a single hydrogen atom as well as a 10 megaDalton atom without changing anything in the system.
Also, semiconductor fabrication techniques should be much cheaper than other MS techniques, so I think that would be an advantage.
Also, one important distinction between our system [and] most of the current MS systems is that our devices are sensitive only to the mass of the molecule and not m/z. So we could detect molecules without ionizing them.
But then the question is: How do you produce and/or transport molecules in vapor form without ionizing?
Were there any specific bottlenecks that you and your colleagues were trying to address as you were developing this technology? Some of the traditional bottlenecks have been low-abundance proteins and the dynamic range of some fluids. Were these things that you had in mind?
We didn't have anything in particular in mind, but it's a new technique that we wanted to address in the high-throughput part of proteomics, where you need a throughput measurement. I think that's where this could be useful.
Also, in terms of the dynamic range of proteins, I think since we're doing quantitative measurements, it could be useful. … You could detect the proteins [that] are in very small [quantities in a sample].
It sounds like the sensitivity is one advantage of the technology. But could it be too sensitive and pick up false-positives? Do you have any data in terms of the false-positive detection rate?
Right now, not really. The way we do it is any frequency shifts that are about two or three times the noise level are automatically rejected. That's what we call false-positive. … But as we go down to single-Dalton sensitivity, I guess there could be other problems that we haven't seen yet.
Would this technology be more useful for targeted analysis rather than a shotgun approach?
I don't think that's true. You could do exactly the things that you do with all the current mass specs with these [NEMS] devices.
[ pagebreak ]
What about in terms of validation? That's one of the next steps in proteomics. It sounds like if this technology works the way you hope it will, that it will enable proteomics researchers to validate their initial findings.
This is just a proof-of-principle experiment, so the next step would be to do this in a tabletop experiment. We haven't started looking at things like validation, or even chemical noise, yet. I guess once we get to a performance level where our system can be compared to some of the other systems that are out there, we will start looking at all these things.
But at this point, our emphasis is more on how we [can] bring the system up to the efficiency and sensitivity levels of some of the other devices out there.
Describe what you're working on as you're optimizing this platform.
At this point … we need about 500 molecules to land on the NEMS before we can tell what the composition is, what the mass of that protein is. This is because the frequency change is proportional to both [the] mass of the protein molecule as well as the position along the NEMS device where it landed. In the next generation of experiments, and we've already started these measurements, what we'll do is something called a multi-mode experiment where we track resonance frequencies called multiple modes. … In these experiments each molecule would produce a distinct frequency shift for each mode and gives us a way to deconvolve the mass.
So instead of waiting for 500 of these molecules to land on the NEMS, we can tell every mass of every single molecule. So that's something we are currently working on.
Also, we've reduced the size of the system from about 2 meters long to a couple of feet, and it's a benchtop experiment right now.
We are also working on the mass sensitivity of the device. In the experiment, it was about 10 kiloDaltons. We are trying to make devices that are slightly more sensitive.
What do you want to bring it down to?
A single Dalton would be good. … We started these experiments, I think, about 10 years ago … at about 10 megaDaltons, but we are now at about 10 kiloDaltons. There are other groups … who've achieved even smaller mass sensitivities.
As you continue developing this, how much of this is applying current nanotechnology methods specifically to mass spectrometry, and how much of it is trying to expand the limitations of nanotechnology in general?
At this point, we are more from the field of nanotechnology, so we try to improve things from that perspective. But once we get to the point where we can say, 'OK, this is good enough, and this is compatible enough with other MS [instruments],' we could collaborate with some of the analytical chemists or some of the mass spec people who could help us do the validation, or help us solve some of the problems that are specific to mass specs.
Are there any specific demands for sample preparation for this device?
Not that I know … we haven't seen any specific very demanding restrictions. … It doesn't require very sophisticated equipment and all of these are compatible with standard semiconductor fabrication techniques.
So fabrication of these devices should be fairly easy.
So in the future, you don't see any issues of manufacturing these devices on an industrial scale?
We don't imagine there will be any problems, and we are collaborating with other people to have these devices produced en masse.
We have an alliance for nano VLSI [very large scale integration]. It's called Alliance for Nanosystems VLSI. …it's a collaboration between Cal Tech and Leti Minatec [a government lab] in France where we are concentrating on making these arrays of devices and making new kinds of devices to reach single-Dalton resolution.
What about mass spec vendors? Have any of them expressed any interest in your technology?
No, not yet, but hopefully with this work, it puts us out there.
As you're developing this, are you developing this specifically for protein research or can this be used for other applications such as metabolomics or food safety?
It could be useful in anything that currently uses a mass spec. We don't see any reason it would not be applicable … so I think it will be useful not only in proteomics, but also if you wanted to detect some other biomolecule, it should be fairly easy to do.