Proteomics.

Following the development of super-sensitive MPD enhanced immunoassays, BioTraces continues the development of improved methods for protein analysis. Two new technologies are under development:

1. MPD enabled proteomics;
2. MPD enabled protein chips.

MPD enabled Proteomics: Current approaches in drug discovery focus on nucleic acids, mostly DNA. But most drugs intervene at the protein level, as secreted proteins or as small molecule inhibitors of proteins. Genomics is producing thousands of raw and unvalidated targets for drug discovery. There is a specific need for a technology that can place genes within their functional context and determine which of them have any relevance to human disease and, of that subset, which genes are apporpriate targets for the development of drugs.

Using its proprietary MPD technology, BioTraces is developing an integrated proteomics system that is at least one-hundred times more sensitive than current techniques. This proteomics system capitalizes on MPD's exquisite instrumental sensitivity. The system enables the highest resolution protein separation, rarest-protein identification, and high throughput. It is best termed a discovery proteomics system, to differentiate it from current-art systems which can identify only the more abundant proteins and are, at best, tools for partial characterization of known proteins. The system will be applied to the entire field of the life sciences, including both human, e.g. drug discovery and plant studies.

The emerging field of proteomics -short for PROTEins analysis using information provided by genOMICS- is the natural extension of genomics. By analyzing patterns of protein expression, proteomics seeks to correlate disease states with specific protein expression and, by extension, gene expression. Genomic databases allow for accelerated protein identification. But it is proteomics which will provide the context to understand a protein's function. An increasing number of companies use proteomics as a tool for drug discovery. These companies use different itterations of three basic technologies: two-dimensional gel electrophoresis for protein separation, mass spectrometry (MS) for protein characterization, and/or Edman degradation for protein identification. Sensitivity of the detection technology is key in proteomics as proteins, unlike DNA, cannot be amplified directly. Thus, only the most abundant proteins are reliably detected. Furthermore, current technologies poorly enable the serial steps of protein separation and identification that are key to proteomics. In practice, only 10-20% of proteins in sample are detected, of which typically less than 25% can be identified and further analyzed. The system includes:

1. Ultrasensitive quantation of high resolution 2D gel electrophoresis using the MPD-Imager;
2. Single gel differential-display of proteins, allowing comparison of two states e.g. diseased versus non-diseased (dd-PROT);
3. Improved artifact elimination and background rejection by novel wavelet image analysis software;
4. High-throughput correlation to sequence via ultrasensitive amino-acid composition;
5. Higher sensitivity MS for peptide mass fingerprints and peptide sequence tagging;
6. MPD enabled ultrasensitive protein sequencing via Edman degradation (Edman/MPD);
7. MPD enhance ultrasensitive functional characterization via DNA and Protein Tool-Kits e.g.ligand/receptor binding studies, DNA protein binding assays, cellular assays.

This integrated system will allow BioTraces to determine differential protein expression patterns relevant to specific disease in a sophisticated and rapid fashion. We will do so on a single gel and down to the smallest protein concentrations. Subsequently, we will sequence or otherwise identify the proteins of interest, and correlate this sequence with genomic databases. We will further characterize high-priority genes/proteins via ultrasensitive DNA and protein assays. The combination of several innovative techniques based on MPD, enables the efficient creation of targeted, high value-added functional databases. The commercial value of such databases may supersede that of current gene expression databases in which entries are either DNA sequences or expressed sequence tags to which no function or corresponding protein product is assigned. A high proportion of protein entries (especially rarely-expressed proteins) will be found only in MPD-derived databases.

MPD Instrumentation for Proteomics: The MPD Imager is supported by proprietary imaging software, Laner for Windows™, that is specifically designed for analysis of blots and gels. We have demonstrated experimentally that the MPD Imager system is about 100 times more sensitive than a phosphor imager.

For proteomics, the most important tool is the MPD Imager; in the last year we improved the performance of our MPD-Imagers by about nine fold. The parameters of MPD Imagers are shown in Table 1. Note, the outstanding sensitivity and fast progress in the number of concurrently measured spots. We have extensively tested MPD Imagers with 2-D protein gels. Spots containing sub-attomoles of 125I-labeled proteins have been reliably quantitated. For about 10 attomole of proteins, the signal-to-background in traditional methods is close to one, while with MPD the signal-to-background is a few hundred. For example, the background in the MPD Imager/121 is 1.5 and 6 zeptomole/mm² for ¹²⁵I and ¹³¹I, respectively. We also demonstrated MPD linearity over more than seven orders of magnitude for samples from picomole to sub-attomole levels. With the MPD Imager/225 overnight quantitation of gels at attomole sensitivity is possible. The MPD Imager/1.6K, which features 1,600 independent scintillator crystals, will permit about seven times higher throughput than the MPD Imager/225. With the MPD Imager/1.6K, large 2-D gels will be quantitated down to the sub-attomole level in a few hours. This limit of detection does not force a compromise between resolution and sensitivity. In fact, the MPD Imager will increase resolution by using lower concentrations of total protein per gel while still allowing a one hundred-fold higher sensitivity than current methods. MPD may allow complete cartography of the proteome, extending from the most abundant of proteins to the most rare.

Table 1. Progress in Development of MPD Imagers

The group of Dr. A. K. Drukier has produced about twenty MPD instruments, of which about ten are currently used in several biological laboratories including NIH, NIST, NASA, NeXstar, Co. and CEB, France.

MPD enhanced protein separation: We developed an MPD Imager which for the 2D distribution of sub-attomole (10-20 mole) amounts of radiolabeled proteins permits a spatial resolution of less than 0.5 mm.

Protein samples with as little as 0.1 pCi activity can be reliably quantitated.

We have estimated the sensitivity and spatial resolution of the MPD-Imager using a ladder of proteins. We produced and studied gels loaded with 1,000 dpm/gel, 100 dpm/gel and 10 dpm/gel, respectively. The ladder consits of 5 proteins with a dynamic range of about a factor of ten. For 100 dpm gel all proteins (including some isoforms) are correctly imaged even if the weakest spots are a few attomoles per spot. The MPD Imager correctly visualized a gel containing a total of less than 10 dpm/gel or less than 50 fg of protein. Using these gels we developed the methods of data processing leading to optimal signal/background ratio.

In collaboration with the group of Dr. H. Langen, Hoffman-LaRoche, we have compared the sensitivity and spatial resolution of MPD-Imager/121 and phosphor imager using E. coli proteins labeled with ¹²⁵I. In the first series of experiments, six gels starting from the same culture of E. coli were produced. There were two triplicates with different amount of proteins loaded per gel. We have demonstrated that MPD Imagers can image low abundance samples that had been labeled with ¹²⁵I. Aliquots of proteins from this sample, equivalent to 0.1 and 0.01 microCi of ¹²⁵I per gel were imaged. The high quality 2D gel shows a few hundred, mostly very low abundance spots (see Figure 1). In these experiments we reliably detected and quantitated even the 0.1 attomole spots, for comparison the limit of detection of silver stain is about 100 femtomole. We present only those spots for which there are more than 10 counts per spot. For 0.1 attomole/spot, the signal/instrumental background ratio is above 50 and the signal/biological background ratio is better than 10. Thus, Figure 1 shows only those spots where the probability of a false positive or false negative is lower than 5%. We estimate that of the approximately 1000 spots visible on this gel only about 10-20 are artifacts. This conclusion has been reached after comparing the gels in triplicate. Furthermore, we "zoomed" into an area of the gel where the weakest spots are present. An example of such a zoomed image is presented in Figure 2a and Figure 2b.Comparison of the "overall" and "zoomed" images confirms that the E. coligel is clear of artifacts, even at sub-attomole levels of protein.

In a second series of experiments, we studied the influence of overloading on spatial resolution. We compared (in triplicate) E. coligels with total loading of labeled proteins of 100 ng/gel and 10 ng/gel. The estimated labeling efficiency (the ratio of labeled to nonlabeled protein molecules) is 0.1. Our data show that with a nominal 10 ng/gel loading, the biological background on the gels is below 10 zeptomole/mm². It is about five times higher for a nominal loading of 100 ng/gel. We also documented, that the lower loading leads to about a factor of three improvement in spatial resolution, probably due to the diminished effects of global overloading. Only spots containing less than 50 attomole (say less than 500 femtogram) of protein have the expected " gaussian" shape. The higher abundance spots have a "hat" shape due to local overloading and the spatial resolution is considerably decreased.

We imaged and analyzed 2D-gels of human serum produced in the laboratory of of Prof. A. Nordheim, Tubingen University. A comparison of the images obtained with MPD and phosphor imagers confirms that a much better signal/background ratio is obtained with MPD Imaging. Many protein spots that were not visible with the phosphor imager could be detected with the MPD imager. For one patient, in the area of interest in the gel [5 cm x 5cm], the phosphor imager detected only 5 proteins whereas the MPD Imager documented reliably the presence of more than 100 proteins (see Figures 3a and 3b). The analysis of the data for this 5cm x 5cm area of interest permitted the enumeration of the well resolved spots (putative proteins). We differentiate between the weak spots (less than 3 pixels) and extended spots (4 or more adjacent pixels). The following Table provides the number of well defined spots.

Table 2.

Thus, when the detection of low abundance proteins is the goal, the MPD Imager performs much better than prior art instruments.

MPD enabled Differential Display of Proteins: Because different isotopes emit photons with different decay signatures, MPD can identify and distinguish different isotopes in the same sample. This capability, which we refer to as multicolor, makes it possible to run multiple assays on the same gel simultaneously. A two color system involving ¹²⁵I and ¹³¹ I has been fully implemented. Multi-color capability permits single-gel differential display. In brief, protein extracts from two samples of interest, e.g. cancer cells vs normal cells, are labeled separately with different tracer isotopes, for example, 125I for cancer cells, and 131I for normal cells. The two extracts are mixed and analyzed on a single 2-D gel. Spots which display differential protein expression and/or post-translational modifications can be detected and analyzed further. This simultaneous analysis contrasts with existing methods which rely on a complex and unreliable multi-gel analysis.

We have demonstrated that MPD Imagers can differentiate co-migrating samples that had been differentially labeled with ¹²⁵I and ¹³¹I in a series of experiments in collaboration with Hoffman-LaRoche. The cross-talk was 0.1% for sample activities down to 0.1 attomole. Experimental results indicate that MPD Imagers permit differentiation between two different comigrating biomolecules. Using E. coli and H. influenzawe were able to differentiate co-migrating samples which had been differentially labeled with ¹²⁵I and ¹³¹I. The crosstalk was 0.1% with sample concentrations down to 0.1 attomoles. More specifically, in a first series of experiments a culture of E. coli was aliquoted and a protein extract was labeled with ¹²⁵I. The same culture of E. coliwas then heated which should lead to expression of heat-shock proteins. The extract of proteins from heated E. coliwas radiolabeled with ¹³¹I. The aliquots of proteins from both samples, equivalent to 0.04 microCi of ¹²⁵I label and 0.04 microCi of ¹³¹I, respectively were mixed. The high quality 2D gel shows a few hundreds, mostly very low abundance, spots and some proteins that seem to be expressed differentially. Control gels with only ¹²⁵I and ¹³¹I were also obtained and imaged. These preliminary results using ¹²⁵I and ¹³¹I labels are highly encouraging and document the advantages of differential display of proteins.

In second series of experiments, an extract of proteins from H. influenza was labeled with ¹²⁵I. The same culture of H. influenzawas then grown in the presence of antibiotics to induce the expression of "SOS" proteins. The extract of proteins from H. influenza in the presence of antibiotics was radiolabeled with ¹³¹I. Aliquots of proteins from both samples, equivalent to 0.1 microCi of ¹²⁵I and 0.03 microCi of ¹³¹I, respectively, were mixed. In total, we studied six gels, one with ¹²⁵I only, one with ¹³¹I and four with both ¹²⁵I and ¹³¹I labeled material. The high quality 2D gel shows a few hundred, mostly very low abundance spots. The differential display permits the detection of about 15 proteins whose abundance is diminished by application of antibiotics (see Figure 4a). Almost all of these proteins are high abundance, house-keeping proteins. We also observed about 50 proteins which are over-expressed when antibiotic is applied (see Figure 4b). Notably, many of these proteins are either not observed or are present only in small amounts in a gel when the antibiotics are not applied. Note, that we compared all four gels and elucidated the common pattern; the same under- and over-expressed proteins appears in all four gels. Identification of the differentially displayed proteins by mass spectrometry is currently underway.

Other topics in protein analysis.

Protein chips: It is increasingly important to develop new techniques which permit monitoring of the metabolic status and cellular response to external stimulation. More specifically, following considerable progress in the Human Genome project, transcription and translation are becoming the frontier of genomic research. We are developing new methods for studying post-translational modifications and their influence on protein secretion. Our method correlates the in-cell and serum abundances. The concurrent quantification for a large number of targets is a major advantage of the proposed super-sensitive, dedicated protein chip concept. We note, that by use of aptamers (i.e., nucleic acid constructs), a transfer of technology from the fast developing microarrays of DNA oligonucleotides is accomplished.

For example, our studies of the levels of several cytokines in AML patients, suggest that the abundances of cytokines are significantly changed in cancer patients. Actually, we documented the first example of down-regulation of interleukin in a cancer case. For the efficient extension of this study to a large number of interleukin, protein chips are required. They will permit us to obtain the differential display of proteins (immuno-profile) of the patients. Thus, our effort has two important justifications: the development of a new techniques enabling ultra-sensitive protein chips, and the elucidation of the correlation of post-translational modulation with disease including better prognosis and therapy monitoring of oncologic cases.

Three elements are important for the successful implementation of the proposed project:

1. aptamers for their specificity and ability to generate the arrays of capture elements;
2. MPD-instrumentation for sensitivity and multicolor;
3. proprietary methods of non-specific biological
background (NSBB) rejection.

We are developing the methodology of differential display of proteins (dd-PROT) both using gels (see above) and using the protein chip concepts. Note, that herein we talk of differential display of known proteins, whose structure and function have been previously elucidated in healthy and sick individuals. For example, an important motivation for the isolation of up or down regulated proteins is to ascertain whether they drive or are correlated with a particular process of differentiation or carcinogenesis. The inadequate sensitivity of current detection methods makes the study of of low abundance protein in physiological fluids very difficult. Also, current methods are not sensitive enough to reliably quantitate the proteins when needle puncture are used to obtain the relatively small number of cells from a particular tissue. MPD based biological methods provide the necessary sub-attomole per sample sensitivity and may enable us to elucidate the presence of translational modulation.

This research is a collaboration of the group of Dr. A.K. Drukier, BioTraces, with the group of Dr. L. Gold, Somalogic Inc. The proprietary SELEX process developed by Somalogic Inc., Pharmaceutical Inc., Boulder, CO, can be used to screen a pool of up to 10¹⁵oligonucleotides and to select for compounds that have the highest affinity to a specific molecular target, e.g., protein. SELEX enables the construction of a library of oligonucleotides (aptamers) with a specificity comparable to, or in some cases higher than, the monoclonal antibodies. Somalogic Inc. has demonstrated that the SELEX process is applicable to a broad range of molecular targets, including proteins, peptides, organic compounds, lipids, polysaccharides, carbohydrates, glycoproteins, and whole cells. The use of aptamers instead of, or in combination with, naturally grown antibodies may permit considerable progress in immunodiagnostics, especially in a quantitation of a large number of protein targets using miniaturized formats ("protein chips").

The combination of MPD instrumentation ("sensitivity") and aptamers ("specificity") may revolutionize immunodiagnostics and protein quantification. Thus, BioTraces and Somalogic Inc. have initiated a joint effort towards the improved detection and quantitation of proteins, viruses and bacteria. The initial goal is to test MPD/Aptamer enhanced methods for detection of cytokine targets and compare their performance with prior-art techniques.