The Human Proteome Organization is nearly finished analyzing the first sample set of proteins in its effort to develop a protein-standard mixture, a project that, based on its current pace, will be completed more than one year later than expected.
In December, HUPO sent out its first sample set of 20 proteins to 24 laboratories worldwide for analysis. So far, about 20 have returned their results. As HUPO waits for the remaining results, the organization anticipates it will send out a second sample set during the summer.
However, HUPO doesn’t yet know what those proteins will be, and it hasn’t chosen the labs that will test them, John Bergeron, immediate past president of HUPO, told ProteoMonitor last week.
In July, HUPO officials said that the entire endeavor would be finished in time for HUPO’s World Congress in Long Beach, Calif., last November [See PM, 07/20/06]. But now, Bergeron says that analysis results from the first sample set will be presented at this year’s World Congress in Seoul in October, and could not say when results from the second set would be known.
“The putting together of the protein mixture at four different sets of varying concentrations for each of the 20 proteins should be soon and so it may be done quickly,” Bergeron said. “I just know that the single mixture will certainly be done by Seoul.”
Bergeron is chairing the HUPO committee overseeing the effort.
For its first sample set, HUPO chose 20 proteins from among 96 and sent them in equimolar amounts to participating laboratories. Bergeron declined to identify the 20 proteins, but said they each have an initial purity level of at least 95 percent with a goal of 99 percent purity within 12 months, according to criteria created by the committee.
Members of the committee include individuals from funding agencies — such as the National Institutes of Health, the Canadian Institutes of Health Research, and the European Commission — scientists, and editors from proteomics journals.
Invitrogen is creating the proteins for both sample sets. Officials from the company were not available for comment, but according to a poster presented by Invitrogen researchers at HUPO’s World Congress in November, the authors identified subsets of the mammalian gene collection of full-length human cDNA clones possessing specific protein and tryptic peptide characteristics to assemble their protein mixtures.
“By starting with a collection of sequence-validated open reading frames, we have avoided the issues of allelic variation that arise when proteins isolated from natural sources are used,” the authors wrote in their poster.
The proteins were expressed in E. coli under conditions designed to maximize the production of inclusion bodies. The researchers used purified inclusion bodies as a starting point to develop protein purification protocols, and used mass spectrometry to identify contaminating E. coli host proteins in semi-purified fractions.
“By purifying and validating individual human proteins, custom blending of specific protein mixes can be performed,” the authors wrote.
Each laboratory received one vial containing five picomoles of each of the 20 proteins, for a total of 100 picomoles. The samples contained four proteins in each of four different molecular weight ranges: 33-36 kiloDaltons; 50-53 KDa; 70-75 kDa; and 100-115 kDa.
The remaining four proteins were distributed among at least two of the four molecular weight ranges.
In addition, each laboratory was sent a “frozen” database — a database that will not change with time due to additions, deletions, or corrections — to ensure that there would be no confusion when the proteins, created by different techniques from various databases, are characterized and matched.
In instances where a laboratory has trouble characterizing certain proteins, HUPO will conduct its own analysis to try to determine what the issue is. If it’s found that the issue is platform-related, HUPO will then send the laboratory another sample so that it can “now be able to characterize all of the proteins to exactly the same degree of comprehensiveness as the so-called very best labs,” Bergeron said.
HUPO has not yet chosen the proteins to be included in the second sample set, though they are expected to be different from the first sample set, Bergeron said. He told ProteoMonitor in July that it will contain 20 proteins in four different samples at varying concentrations spanning either three or four logs with some kept constant at high abundance, some at low abundance, and others mixed in.
HUPO also hasn’t chosen the labs that will test the second set.
Invitrogen has the rights to commercialize the protein standard mixtures. In the fall, Sigma Aldrich launched the first commercially available mixture, based on a study done by the Association of Biomolecular Resource Facilities on protein standard mixtures [See PM 11/02/06].
Behind Schedule — Predictably
HUPO is already more than five months behind schedule in unveiling results from its protein standard mixture initiative. Now it appears its study may not be completed until the end of the year.
“[With] any high-throughput, to come up with decent reference sets [of standards] is very tough. And it’s tough because of technological issues, because of experimental design issues, [and] money, of course.”
Based on the efforts of others who have developed protein-standard mixtures, the delay is not a complete surprise. Developing protein standards is a tricky business, they said, and setbacks and complications are inevitable.
With any high-throughput methodology, coming up with “decent reference sets” of standards is “very tough,” said Eugene Kolker, a HUPO member involved in the protein standard-mixture initiative. “And it’s tough because of technological issues, because of experimental design issues, [and] money, of course,” said Kolker, who is also president and director of the Biatech Institute, a nonprofit that has developed and distributed two protein-standard mixtures of its own and is currently working on a third.
A year ago, as ABRF was doing its protein standard mixture study, it ran into several unexpected roadblocks, said Jeffrey Kowalak, a staff scientist at the National Institute of Mental Health, and member of the ABRF research group that conducted the study.
ABRF had set as a minimum requirement a protein purity level of 95 percent, but during pilot studies, he said, it became apparent that about 40 percent of the proteins being created would not be able to reach that level, and so had to be discarded from the final mixture.
In addition, during analysis of the protein mixture, the laboratories discovered 87 “bonus proteins,” or proteins that were not intentionally included in the study’s protein mixture.
Further informatics analysis showed sequence similarities among some of those 87 proteins, reducing the list of “bonus proteins” to about 20, Kowalak said. It remains unclear what caused them to appear in the sample set, though unintentional contamination and or artifacts generated in the informatics stage are the two most suspected causes, he said.
“It’s very, very difficult to formulate a standard of intentionally known composition, even when you’re working with highly purified proteins,” he said. “There are always other proteins that adventitiously copurify.”
Even after getting the proteins to the required minimum 95-percent purity level, HUPO found that non-science related issues, such as getting the samples past customs inspections, presented hurdles that needed to be overcome.
“Just getting the samples to [the laboratories] proved to be a little more challenging than we thought,” Bergeron said.
In spite of the delays, Bergeron said the goals of the protein-standard initiative haven’t changed.
“We work with the laboratories in order to figure out in their pipelines, in an educational way, how the characterization can be done in a very productive, facile way,” he said.
HUPO’s goal “remains the same: to assure the proteomics community, through HUPO, that a comprehensive, accurate and permanent proteome can be realized in a sample with modern proteomics-based platforms,” he said.