Making FT/MS Matter

A few proteomics researchers have demonstrated the power of Fourier transform mass spectrometry, and now vendors are trying to make their million-dollar instruments mainstream

By John S. MacNeil

Like many protein mass spectrometrists, Andrej Shevchenko, at the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden, is on the fence about Fourier transform mass spectrometry. With a lab group of two post-docs and three students, he’s not in a position to plop down upwards of $1 million to acquire a high-magnetic-field-strength spectrometer, nor is he willing to commit his group to customizing an FT/MS instrument for biology as his primary research focus. “If I cannot be the first, it doesn’t pay off to be the second or third and have the same technical problems [as the pioneers of the technology],” he says.

A combination of factors is changing this perception, however. In recent years, as FT/MS specialists have begun to demonstrate the power of their tool of choice for proteomics, more biologists are paying attention. At the American Society for Mass Spectrometry meeting a year ago, FT/MS approaches to proteomics stood out as one of the highlights of the conference for their superior results, and this year’s meeting this month in Montreal is sure to offer more of the same. Aware of these advances, mass spectrometry vendors have stepped up to the challenge of adapting FT/MS for more mainstream and user-friendly applications. At this year’s Pittcon meeting in Orlando, both Bruker Daltonics and Thermo Finnigan unveiled new hybrid FT/MS instruments that promise to make the high-performance mass spectrometry technique less expensive and easier to use for many biological mass spectrometrists looking to identify the components of complex protein mixtures.

The impact of more researchers using FT/MS for proteomics could be significant. Early last year Fred McLafferty, an FT/MS pioneer at Cornell University, showed that his version of hybrid FT/MS could identify the type and location of almost every post-translational modification to a 29 kDa protein. And in August of last year Dick Smith, who runs a large FT/MS group at Pacific Northwest National Lab in Richland, Wash., completed the most thorough analysis of any organism’s proteome to date by identifying 61 percent of the predicted proteins in Deinococcus radiodurans, a radiation-resistant bacterium. “Where interfaced with biology, FT/MS would be very powerful,” says Ruedi Aebersold, a proteomics expert and co-founder of the Institute for Systems Biology in Seattle.

There’s a lot of excitement around these new tools for big biology, but Shevchenko has seen it before: “People tend to believe in something which is promising but where nothing has been yet achieved. The best results are always the next results, not the results from yesterday. So if we haven’t succeeded in FT/MS then FT/MS is the way to go,” he says.

The question for the new hybrids is, will the boost they provide to protein analysis offset their cost? Or will the initial enthusiasm for FT/MS fizzle out — in the manner of the MALDI-TOF/TOF mass spectrometer, another much-heralded tool for proteomics? It may be too early to tell, particularly since Thermo’s new hybrid FT/MS instrument won’t be ready to ship until September.

But Bruker already has a few customers. David Muddiman at the Mayo Medical School in Rochester, Minn., has installed a version of Bruker’s hybrid quadrupole FT/MS instrument, which required technicians to open a hole in the roof of the laboratory to drop in the 12-Tesla superconducting magnet. The University of Amsterdam has also purchased a hybrid system from Bruker, and Sandia National Laboratory has decided to upgrade its current FT/MS system to the hybrid version, Bruker says. Bruker’s hybrid FT/MS instrument ranges in price from $1.2 million to $2 million, and Thermo’s hybrid ion-trap FT/MS will go for about $750,000 — primarily because the instrument comes with a weaker magnet.

How Fourier Transformed Proteomics Posthumously

FT/MS is widely accepted as the paragon for performance in mass spectrometry. Its unbeatable ability to calculate the masses of ionized molecules and to distinguish between two ions similar in mass are attributable in part to the high magnetic field required to operate the instrument.

But the FT/MS’s power also relies on a theory that dates back to the 19th century. In the early 1800s Joseph Fourier, a French mathematician who had studied under Lagrange and Laplace, decided to tackle the problem of how to mathematically describe heat propagation. The reasons for Fourier’s choice of study are unclear, but he had served as Napoleon’s science advisor during the French occupation of Egypt from 1798 to 1801 and, according to some accounts, had become obsessed with heat, often keeping his rooms uncomfortably hot.

Later, while serving as an administrator for the French province of Isèe, Fourier developed a theory of heat conduction that relied on the idea that the rate of heat transfer was dependent on the difference in temperature between two objects and their distance from each other. (Most other scientists’ analyses at the time focused on the idea of an object having an inherent property called caloric, which determined its temperature). Fourier also proposed that one could mathematically extract frequency information from an equation describing oscillating behavior. It wasn’t until 165 years later, however, that scientists figured out how to use the formula to make detecting the presence of particular ions in a mass spectrometer dramatically more efficient.

A Brief History

In 1973, Alan Marshall and his colleague Melvin Comisarow, then both junior professors at the University of British Columbia in Vancouver, were working with a form of mass spectrometry called ion cyclotron resonance, in which ions rotating in the presence of a magnetic field are individually probed to determine how fast they are rotating. These frequencies are inversely proportional to the ion’s mass-to-charge ratio, which in turn can be used to calculate the ion’s mass. However, because researchers had to test for each ion’s presence separately, the technique was slow and sample-consuming, and mass resolution was limited by the need to scan the magnetic field. Marshall and Comisarow came up with a better idea: why not probe all the ions at once, and then use Fourier transform analysis to pull out all the constituent frequencies? The idea worked quite well, in fact, and Fourier transform ICR mass spectrometry was born. Since the resolving power of the FT-ICR system increases with the strength of the magnetic field, given a strong enough magnet, researchers could outdo all other forms of mass spectrometry in resolution and mass accuracy — a fact that remains true today.

Extremely high-strength magnets, not surprisingly, are extremely expensive. For almost 25 years FT-ICR mass spectrometry remained the domain of specialists devoted to improving its capabilities. Until recently, scientists hadn’t figured out how to ionize biological compounds, which tend to fall apart under the heat-intensive methods then used to break molecules into their charged components. It wasn’t until the 1980s that researchers such as John Fenn at Virginia Commonwealth University and Franz Hillenkamp and Michael Karas at the University of Münster developed the electrospray ionization and matrix-assisted laser desorption ionization techniques, known as ESI and MALDI, that allow mass spectrometers to analyze large biological molecules like proteins.

The tide was turned by a handful of mass spec gurus led by Marshall, McLafferty at Cornell, Don Hunt at the University of Virginia, and Bob McIver, then at the University of California, Irvine. At the time, the problem with applying FT/MS to proteins lay in the difficult task of fragmenting large ions, known as parent ions, into smaller pieces to identify the individual peptides constituting the polypeptide chain of the protein. Complicating that problem was the issue of controlling the number of ions entering the FT-ICR cell, which operates under high vacuum at pressures below 10-9 torr.

The solution, Hunt and McIver determined in the mid-’80s, was to construct a hybrid mass spectrometer. Hunt and his team tacked an ion-trap analyzer to the front end of their FT-ICR, initially just to focus ions so that they would pass through the fringe field of the magnet into the ICR cell. At Cornell, McLafferty’s group decided to slap on a series of quadrupole analyzers to the front end of the FT-ICR analyzer, a configuration that allows a researcher to select only certain parent ions for further fragmentation before introducing them into the high-vacuum chamber.

“Taking advantage of the unique characteristics of each mass analyzer and combining them makes an instrument that is particularly powerful,” says Hunt. “When you do that with this new instrument it turns out what you get is beautiful fragmentation patterns, a higher dynamic range than you can get on the ion trap per se, accurate mass measurements and high resolution that you can’t get on the ion traps, and because there’s no chemical noise the sensitivity goes up by more than a factor of a thousand.”

PNNL: FT/MS Matters for Proteomics

The scientist who could be said to have done the most to demonstrate the power of FT/MS for proteomics is Dick Smith, a chief scientist and protein mass spectrometrist at PNNL. In 1991, as part of planning the new $200 million Environmental Molecular Sciences Laboratory, he spearheaded an effort to build an $8 million, state-of-the-art FT-ICR mass spectrometry facility that would address the Department of Energy’s interest in fundamental biology.

Smith set about optimizing his FT-ICR for proteomics by converting the system into a hybrid quadrupole/FT-ICR and automating the front-end capillary liquid chromatography and electrophoresis systems for separating complex mixtures of proteins and peptides. He also designed an electrodynamic ion funnel to improve the instrument’s sensitivity. By 2002 he could boast that his lab had completed the most thorough analysis of any organism’s proteome to date, in which he identified 61 percent of the predicted proteins in Deinococcus (that number now stands at 83 percent). In addition, Smith led the development of ways to make extremely accurate measurements of peptide masses, and he devised a new mass spectrometry-based approach for removing the most abundant protein ions in the mass spectrum from the sample, called DREAMS (dynamic range enhancement applied to mass spectrometry).

Smith’s advances have not gone unnoticed — by the scientific community or by instrument manufacturers. In 2001, Bruker Daltonics signed a partnership agreement to commercialize a version of Smith’s hybrid FT/MS (Bruker and IonSpec had for many years sold a conventional FT/MS instrument primarily for small-molecule analysis). Smith had purchased his initial instruments from Bruker, so the company latched onto the PNNL group as development partner. “Their group has pioneered FT/MS proteomics, and we believe the direction they were heading in and are still heading now is most likely the way to go with FT/MS,” says Mike Easterling, FT/MS applications manager for Bruker. At this year’s Pittcon, Bruker launched its Apex-Q hybrid quadrupole FT/MS instrument.

Earlier, in 1997, Thermo Finnigan had decided to commercialize its own version of a hybrid FT/MS instrument based on Don Hunt’s hybrid ion-trap FT-ICR research spectrometer, says Lester Taylor, Thermo’s global product marketing director for life sciences mass spectrometry. At the time, former members of Hunt’s lab were working with Toronto-headquartered MDS Proteomics to install a version of the hybrid instrument as part of MDSP’s discovery proteomics platform. This year at Pittcon, Thermo promised to begin shipping the spectrometer, known as the LCQ-FT, by September.

Micromass and IonSpec also plan to launch their own version of a hybrid quadrupole FT/MS instrument. Last August, the two companies announced a partnership to build a hybrid instrument combining Micromass’ ESI interface and quadrupole technology with IonSpec’s FT-ICR analyzer. Although Micromass and IonSpec seem to be playing catch-up, IonSpec co-founder Bob McIver was one of the originators of hybrid FT/MS, and holds several patents on the technology. Micromass and IonSpec said last year that they would introduce the new instrument by the first quarter of 2003, but Jeff McIver, the company’s director of marketing, now says the new hybrid is “in development as we speak,” and that the company will have more details on its availability at the American Society for Mass Spectrometry meeting in Montreal this month.

Vendors Jockey for Position

Now, in the run-up to a hybrid FT/MS sales race, vendors are jockeying for position (see table p. 36). Bruker says its Apex-Q, available with a magnet of up to 12 Tesla, provides researchers with a package of options for applying FT/MS to proteomics. Compared to Thermo’s LCQ-FT, which comes equipped with a 7-Tesla magnet, Bruker’s 9.4- and 12-Tesla Apex-Q instruments should offer higher performance for the discerning mass spectrometrist, the company says.

In addition, because the Apex-Q incorporates techniques for fragmenting ions called electron capture dissociation and infrared multiphoton dissociation initially developed by McLafferty at Cornell, Bruker says only its hybrid FT/MS spectrometer can analyze intact protein ions. Known as “top-down” proteomics, the approach allows researchers to glean more specific information about a protein’s post-translational modifications, Bruker says.

The technique works as follows: Rather than chop up a protein into peptides with an enzyme such as trypsin prior to injecting them into the spectrometer, McLafferty and his disciples use ECD to cleave the protein backbone specifically, leaving the bonds between peptides and any carbohydrate or phosphate modifications intact. Although proteomics experts continue to debate the relative merits of top-down proteomics, in recent papers McLafferty’s group and others have demonstrated that the technique has potential for general application. In a paper published in February 2002 in the Proceedings of the National Academy of Sciences, McLafferty’s group showed that it could cleave 250 of the 258 peptide bonds in a 29 kDa protein called carbonic anhydrase, allowing the group to identify the type and location of almost all the protein’s post-translational modifications.

Acknowledging that his is the view of “a very prejudiced person,” McLafferty says, “I’d like to think that this top-down approach is the thing that’s really an important driving force for the [introduction of hybrid FT/MS instruments], in that you no longer need to do this wasteful — it wastes sample and it wastes information — breaking up the protein before you put it in the mass spectrometer.”

Thermo, on the other hand, says it’s targeting a more focused segment of the biological mass spectrometry marketplace. With its 7-Tesla magnet and ion-trap front-end, the LCQ-FT should appeal to proteomics researchers looking for performance advantages over hybrid quadrupole time-of-flight mass spectrometers who also don’t want to spring for the cost of a higher strength 9.4- or 12-Tesla magnet, says Thermo’s Taylor. In fact, Taylor says that Thermo’s 7-Tesla hybrid ion-trap FT-ICR offers comparable performance to higher magnetic field strength hybrid quadrupole FT-ICR.

“The great thing about this instrument is that it defies most of the usual maxims that as you increase resolution you lose sensitivity,” says Taylor. “Sensitivity is essentially governed by the front stage linear trap on the instrument, and because we are ejecting over 90 percent of the ions into the [FT-ICR] cell we don’t lose anything in the transmission.”

In addition to these boasts, Thermo says its LCQ-FT is uniquely suited for routine high-throughput MudPit (MultiDimensional Protein Identification Technology)-type analyses that use front-end liquid chromatography separation schemes and back-end peptide database searches to piece together and identify the proteins present in a sample. And researchers familiar with the instrument say that if the company can hide the complexity of the FT/MS analyzer behind its familiar ion-trap interface, it may have a good chance at converting the customers who value Thermo’s ion traps for their ruggedness and reliability.

“Finnigan has a really good track record of making user-friendly and robust instruments,” says Ruedi Aebersold. “So if their trap FT/MS gets anywhere close to a trap in terms of robustness and ease of use, [combined] with the increase in performance, it will be a very powerful instrument.”

Dreams of Mainstream

Because the first hybrid FT/MS instruments have become available only in the last few months it remains an open question as to whether the technique will cross over into mainstream biological mass spectrometry. Those who have so far purchased Bruker’s Apex-Q, such as David Muddiman at the Mayo Clinic, are not newcomers to the technology. Muddiman, a professor of biochemistry and molecular biology at Mayo Medical School, worked as a post-doc under Dick Smith at PNNL, and his lab at Mayo also holds two IonSpec FT-ICR mass spectrometers.

Directors of academic protein mass spectrometry core facilities say they’re waiting to see more published performance data, but are also intrigued by the advantages that hybrid FT/MS technology might offer their users. “There are indications that FT/MS is about to enter into a new phase,” says Mark Duncan, director of the Biochemical Mass Spectrometry Facility at the University of Colorado Health Sciences Center in Denver. “FT/MS has some remarkable attributes. The resolution and the sensitivity of that instrument make it a powerful component of a hybrid instrument — as the back end — so I think there’s every reason to be very optimistic.”

Bill Lane, director of the Microchemistry and Proteomics Analysis Facility at Harvard University, says he never was too excited about a standalone FT/MS, but is seriously considering the LCQ-FT from Thermo because the platform offers the promise of high sensitivity on the ion-trap front end, and increased resolution and mass accuracy on the FT/MS back end. “In our experience the current hybrid quadrupole-TOF platforms trade what I would call the ‘raw proteomics sensitivity’ that ion traps have — the ability to dig in the dirt for extensive mixture analysis and protein coverage — for their increased mass accuracy and resolution,” Lane wrote to GT in an e-mail. “Our expectation is that a linear ion-trap FT/MS hybrid will give us both without compromise.”

What would really make hybrid FT/MS instruments appealing, mass spectrometrists say, is if vendors could build a machine capable of handling samples at significantly high throughput. A combination of MALDI-TOF and ion trap mass spectrometers may be a better option for researchers trying to identify large numbers of proteins quickly and accurately, says Duncan, but a hybrid FT/MS machine that offered high throughput and ease of use might tilt the scale. “The speed at which you can get fragmentation data is still not there,” says Neil Kelleher, an FT/MS specialist at the University of Illinois at Urbana-Champaign. “This is where FT/MS still falls flat.”

The Biology Challenge

Even if hybrid FT/MS does manage to become a staple of biological mass spectrometry laboratories, there is debate about how much of an impact the technique will have on advancing researchers’ understanding of the biology itself. After all, some say, more data does not necessarily equal deeper insight; much of the value of a high-end instrument like the hybrid FT/MS will come from properly designed experiments with high-quality samples.

From this perspective, proponents of approaches such as top-down proteomics that rely almost solely on FT/MS must justify its importance without resorting to the defense that “anything new is better.” Aebersold at the Institute for Systems Biology says information on post-translational modifications would be useful, but in order to determine the biological signficance of a particular modification, “top-down proteomics would have to be embedded into a wider biological strategy.”

Kelleher at the University of Illinois believes the ability to rapidly characterize post-translational modifications will ultimately find widespread application in pharmaceutical companies, as researchers realize that studying these modifications will help them more rapidly screen through potential drug targets.

“For drug targets that have many different flavors in cells, you have to get all that information, and you will realize that mapping of proteins with complete knowledge of post-translational modifications is important,” Kelleher says. “To the extent that you can’t characterize proteins as efficiently or with enough information from standard bottom-up proteomics, whether you’re using FT/MS or not, you might come to the realization that top-down really does have a role to play.”

Even without resorting to top-down approaches, there are groups that have found applications in drug discovery for FT/MS as it currently exists. One of the most innovative industrial FT/MS scientists is Steven Hofstadler at Ibis Therapeutics, a unit of Isis Pharmaceuticals in Carlsbad, Calif. Although Hofstadler works with standard FT-ICR mass spectrometers from Bruker, he’s developed techniques for applying the technology to screening for drug interactions with RNA molecules. Hofstadler, who also served time as a post-doc in Dick Smith’s lab at PNNL, says his group is the only in the world to apply FT/MS to screening RNA drug targets, and he adds that there’s no reason why the approach wouldn’t also work to screen for drug interactions with proteins.

Although most pharmaceutical companies — if they have an FT/MS instrument — use it for exact mass measurements of small molecules to study compound purity and other quality control issues, that could change if the new hybrid FT-ICR instruments take off. Using his hybrid quadrupole FT-ICR, Alan Marshall at Florida State demonstrated in a recent paper in the Journal of Molecular Biology that hydrogen-deuterium exchange mass spectrometry could help show how the capsid protein for HIV virus is assembled. “That starts developing drug targets. If you tell people where those things stick together then they know where to go to target to try to break them apart.”

The Dilemma Remains

Shevchenko at Max Planck agrees that FT/MS offers unbeatable performance, and that the new hybrid instruments make the technique even more powerful for proteomics. But as a biological mass spectrometrist with a relatively small group, he’s still wary of committing to an approach that might limit his flexibility in taking on different types of problems. For a mass spectrometry core facility, he says, the instrument might make perfect sense, but as to what specific biological problem he’ll be studying, “I don’t really know what I’ll be doing in two years,” he says.

Shevchenko makes the analogy to a car race: “To be competitive at Formula One [racing], I would need a Formula One car. But with a Formula One car, it’s expensive and you need a whole team to maintain it. You can only do Formula One racing; you cannot go and participate in Paris-Dakar. You are basically stuck in one particular application, and you can do that with amazing efficiency but you cannot do anything else.”