EP1060377A2 - Reagents and methods for protein microsequencing - Google Patents

Abstract

The invention provides methods of peptide and protein microsequencing using radiolabeled reagents to perform peptide and protein microsequencing at levels lower than 10 femtomole, preferably at a level lower than 1 femtomole.

Description

REAGENTS AND METHODS FOR PROTEIN MICROSEQUENCING

Field of the Invention

This invention relates to methods and radiolabeled reagents for super-sensitive protein microsequencing and amino acid analysis. More specifically, methods of radiolabeling of Edman- type reagents and their use with multiphoton detection are disclosed.

Background of the Invention

Biomedical applications require ever increasing sensitivity for the characterization of single organisms (virus, bacterium) or single cells. Especially important is the detection of differences in protein composition of normal vs. diseased cells, e.g. cancer cells. Multiphoton Detection enhanced 2D gel separation permits specific selection of a protein from a complicated mixture. The prior art limit of detection (LOD) for proteins is typically on the order of pg/ml or 0.1 fmole/ml. Recently, BioTraces, Inc.'s MPD™ Multi Photon Detection technique achieved 0.01 pg/ml or 0.5 attomole/ml sensitivity. Not only the detection, but also the quantitation of the protein is important, for example in the case of cancer markers. Often, each target should be characterized, e.g. in the case of metastatic cancer cells which have to be distinguished from millions of virtually identical healthy cells. Thus, in biomedical applications the ability to microsequence polypeptides and proteins is of increasing importance. Alas, the prior-art methods of microsequencing require a rather large amount of proteins, typically a few picomole. In many of the above applications, only attomoles of a given protein may be available. Thus, currently, the proteins from hundreds and thousands of cells have to be pooled. This, however, means that in the prior-art only averages and not distributions of biologically important proteins can be measured and characterized.

Currently, microsequencing sensitivity is a limiting factor. At attomole amounts, purification and non specific biological background (NSBB) may become a limiting factor. Microsequencing is a method which also provides considerable redundancy and facilitates NSBB suppression. Methods are required which permit a major improvement (by a factor of a hundred and more) of sensitivity of polypeptides and proteins microsequencing. A supersensitive method for quantitation of radiolabeled, especially radioiodinated polypeptides and proteins, uses MPD™ MultiPhoton Detection (MPD). MPD technology enhances high performance liquid chromatography (HPLC) and permits reliable quantitation of some radiolabeled analytes down to 0.1 attomole/sample.

An important aspect of the MPD™ Multi-Photon Detection technique is its unparalleled capabilities in detection, quantitation and analysis of proteins. In the last few decades, the availability of methods for the in vitro amplification of DNA or RNA has revolutionized the field of biological research and has had a profound impact on the way scientific questions are being addressed. When DNA or RNA molecules are involved, amplification ofthe molecule of interest can be carried out in vitro prior to analyses and/or subsequent manipulations, e.g. cloning, sequencing, in vitro transcription, etc. As a consequence, approaches to biological questions have been considerably biased towards a genomic point of view, according to which no question can be addressed properly unless the genes involved can be cloned. However, the true actors of "life" are not genes or their transcripts, but proteins which are endowed with multiple functions: catalysis of biosynthesis and metabolism, signal transduction, bioassembly, etc. It should be stressed here that knowing the gene or being able to quantify the amounts of transcript that encode a given protein is only a part of the information. This protein may be qualitatively and quantitatively post-transcriptionally modified in such a way that it is either active or inactive, is expressed in either large or small quantities, is either secreted or bound to the membrane etc.

It would be great if another revolution parallel to PCR had occurred in the world of protein analysis. Alas, proteins are not self-replicating template molecules like DNA and there is no way to amplify a protein in vitro other than to isolate and amplify the RNA transcript or gene that encodes it and then translate it in vitro. And even if this tedious procedure is to be used, which is only possible if the gene sequence is known, the copies may not be identical with the original due to post-translational modifications (e.g. chaperon-aided folding, glycosylation etc) that occur in vivo, but are difficult to detect and/or to reproduce in vitro or in different organisms. Therefor, the fundamental issue in protein analysis is the sensitivity of the methods to detect, analyze and manipulate the proteins. Current detection methods have reached their limits of sensitivity which explains why it is easier to study genes, rather than their biologically crucial products, i.e. proteins. The important step is to correlate the genomic content of a given set of cells with the cellular functions, including the expression and secretion of rare proteins.

There is a need for new methods to enable the creation of better protein databases and ways to efficiently correlate them with DNA databases. The improved sensitivity would allow a novel, direct investigation methodology: isolate a protein of interest, characterize its properties (level of expression, binding to a given molecule, catalytic activity) and finally identify it by supersensitive microsequencing. This would permit the identification of the gene which codes for it and mass-production ofthe protein.

SUMMARY OF THE INVENTION

The invention provides methods of peptide and protein microsequencing using radiolabeled reagents to perform peptide and protein microsequencing at levels lower than 10 femtomole, preferably at a level lower than 1 femtomole.

Bioreagents according to the invention are Edman-type reagents which have an aromatic chemical structure, a reporter moiety(moieties) that is or may be derivatized to include a radiolabel, a purity preferably better than 99.99% level, and an ability to react with amino acids by means of an active group such as isothiocyanate. The radiolabels are compatible with MultiPhoton Detection techniques. They may be selected from the family of EC emitters and positron/gamma emitters. They may be halides, including radioisotopes of iodine, including ¹²³I, ¹²⁴I, ¹²⁵I, and ¹²⁶I. They may be EC emitters from the lanthanides family or the platinides family.

The reagent may be radioiodinated Edman reagent. It may be n-[ Ijiodophenyl isothiocyanate, and wherein n=2,3,4,5,6. It may be n,m-[¹²⁵I]diiodophenyl isothiocyanate, and wherein n,m = 2,3 ; 2,4 ; 2,5 ; 2,6 ; 3,4 ; 3,5 ; 3,6. The isothiocyanate group is alternatively a substituent in an aromatic ring other than phenyl such as naphtyl, anthracenyl, pyrenyl, pyridyl and the like. The hydrogen atoms in the phenyl or other aromatic ring are preferably substituted with groups such as methoxy, hydroxy, and the like in order to facilitate radioiodination without significantly diminishing reactivity of the isothiocyanate moiety. The radioiodinated Edman reagent can be radioiodinated before or after the degradation cycle and before or after the separation step. Preferably the radioiodinated Edman reagent used in the degradation has a protected active group that is deprotected and labeled before the separation step. The active group is preferably a primary or secondary amino group.

The Edman-type reagent may be 4-tert-butoxycarbonylaminomethylphenyl isothiocyanate. It may be that the radiolabel is introduced into the Edman-type reagent by means of a linker which itself is attached at the n position of the aromatic ring, wherein n=\, 2,3 ,4,5,6. The linker is preferably an alkyl chain containing 1 to 30 carbon atoms terminated with a reporter group. The reporter group is preferably a primary or secondary amino group or an active group such as aldehyde, hydroxyl, carboxyl, sulfhydryl, phosphate, phosphorothioate, azido, and the like or combination thereof. The linker chain contains for example 1 to 15 heteroatoms, such as oxygen atoms.

Radiolabeling may be achieved using [¹²⁵I]iodinated N-succinimidyl-3-(4-hydroxyphenyl) propionate (Bolton-Hunter reagent), [¹²⁵I]iodinated N-succinimidyl-3-(4-methoxyphenyl) propionate (modified Bolton-Hunter reagent), metal [¹²⁵I]iodide salt or salts for direct iodination in oxidizing conditions, [¹²⁵I]iodinated reagents and any type of chemistry used generally for chemical conjugation.

The excess of the labeling reagent may be removed by simple physicochemical operations such as extraction, gel filtration, gel chromatography, precipitation, adsorption and the like.

The product of the cleavage step (e.g. anilinothiazolinone, ATZ) may be reacted with the primary amine group of a supplemental reagent. In this case, the supplemental reagent may be radiolabeled, while the Edman reagent is unmodified.

The method of the invention may be used for microsequencing using a radiolabeled generalized Edman-type reagent in a procedure comprising the following steps:

* classic Edman degradation based peptide/proteins microsequencing.

* high sensitivity separation. * quantitation ofthe separation process output with means of MPD instrumentation.

The method may be used for amino acid analysis comprising the following steps:

* polypeptide or protein exhaustive hydrolysis;

* derivatization with Edman reagent resulting in PTH-amino acid product;

* high precision measurement of radioactivity ofthe above products;

* high sensitivity separation;

* quantitation ofthe separation process output by means of MPD instrumentation.

The separation process for both embodiements may be for example high performance liquid chromatography (HPLC), thin layer chromatography (TLC), capillary electrophoresis (CE), or gel electrophoresis (GE).

A device according to the invention comprises a spatially resolving MPD (SR-MPD) detector. The SR-MPD is for example initially placed around the HPLC column, typically in the middle ofthe column but is movable and it's movement is computer steered to follow up movement of the peak of analyzed Edman degradation product within a column. The SR- MPD may comprise a series of linearly arranged scintillator crystals each with the dedicated photodiode or avalanche photodiode readout.

The SR-MPD movement is for example initiated after the detector registers at least 10 counts providing that the count rate is compatible with previously measured activity of Edman degradation product. SR-MPD movement may be initiated after the detector registers at least 10 counts providing that the count rate is compatible with previously measured activity of Edman degradation product. The SR-MPD movement is preferably computer controlled, so that the maximum count rate is at the center of the said SR-MPD. SR-MPD signal registration is for example performed several times with increasing resolution, preferably three times.

Resolution improvement can be achieved by the use of apertures of various dimensions. Resolution improvement can be achieved by the use of various step lengths, or a combination of changes in aperture dimensions and step length. For amino acid analysis, the TLC may be TLC on silica gel, TLC on a Reversed Phase carrier. The TLC plate is analyzed by means of MPD instrumentation, preferably an SR- MPD device. The separation process is preferably performed simultaneously for many sequencing reactions in parallel. The number of sequencing reactions performed in parallel may be in the range 1-200.

The migration of the Edman degradation product can be stopped when still inside the HPLC capillary, and an MPD Imager used to quantitate the distribution of radiolabeled Edman degradation product inside the capillary. An SR-MPD can be initially placed around the capillary, typically in the middle ofthe column but is movable and with movement computer steered to follow movement of the peak of analyzed Edman degradation product within the capillary. The SR-MPD may be linear and comprise a series of linearly arranged scintillator crystals each with the dedicated photodiode or avalanche photodiode readout. A linear SR- MPD may have movement initiated after the detector registers at least 10 counts providing that the count rate is at compatible with previously measured activity of Edman degradation product. It may have movement initiated after the detector registers at least 10 counts providing that the count rate is at compatible with previously measured activity of Edman degradation product. The linear SR-MPD movement is preferably computer steered so that the maximum count rate is at the center ofthe said linear SR-MPD.

BRIEF DESCRIPTION OF THE FIGURES

The invention is better understood by reading the following detailed description with reference to the accompanying figures, in which like reference numerals refer to like elements throughout, and in which:

Fig.1 : Edman degradation cycle (prior art)

Fig.2: Linearity of HPLC/MPD measurements of a mixture of steroidal hormones; the results for dilutions down to a factor of 1:1024 are shown. The lowest point corresponds to 0.02 attomole/sample. Fig.3: Reproducibility of HPLC/MPD measurements of a few attomoles/sample of a mixture of estradiol and testosterone; the total count at the peaks is plotted. Fig.4: Well resolved peak of insulin for concentration of 5 attomole/sample. Fig.5 : Linearity of HPLC/MPD for insulin. Fig.6: Reproducibility of HPLC/MPD for insulin. Fig.7: Radioiodination of Edman reagent Fig.8 : Radioiodination of Edman reagent

DETAILED DESCRIPTION OF THE INVENTION

In describing preferred embodiments of the present invention illustrated in the drawings, specific terminology is employed for the sake of clarity. However, the invention is not intended to be limited to the specific terminology so selected, and it is to be understood that each specific element includes all technical equivalents which operate in a similar manner to accomplish a similar purpose.

Abbreviations

Attomole 10^"18 mole = 6 x 10⁵ molecules

APG Aminopropyl glass

APS Aminopolystyrene

ATZ 2-Anilino-5-thiazolinone

BAMPITC Boc-aminomethyl-phenylisothiocyanate

CE Capillary electrophoresis

DABITC 4-N,N-Dimethylaminoazobenzene-4'-phenylisothiocyanate

DCC Dicyclohexyl carbodiimide

DITC p-Phenylenediisothiocyanate

DMAA Dimethylallyl amine

DMBA Dimethybenzyl amine DNSAPITC Dimethylaminonaphthylsulfonylamino-phenylisothiocyanate(dansyl isothiocyanate) EC Electron Capture

Femtomole 10^"15 mole = 6 x 10⁸ molecules FITC Fluorescein-isothiocyanate

FMOC-C1 Fluorenylmethoxycarbonyl chloride FN False negative FP False positive

HFBA Heptafluorobutyric acid

HPLC/MPD™MPD™ enhanced High Performance Liquid Chromatography

LIF Laser-induced fluorescence LOD Limit of detection

MITC Methylisothiocyanate

MMT Monomethoxy trityl (monomethoxytriphenylmethyl)

MPD™ Multi Photon Detection

NITC Naphthyl isothiocyanate NSBB Non-specific biological background

NSL Non-specific losses

OPA O-phthalaldehyde

PITC Phenylisothiocyanate

PTH Phenyl thiohydantoin SR MPD™ Spatially resolving MPD

TETA-resin Tetraethylenetetramine-polystyrene resin

TFA Trifluoroacetic acid

TLC/MPD™ MPD enhanced Thin Layer Chromatography

Zeptomole 10^"21mole=600molecules

Polypeptide or protein as used here means proteins, short peptides, and other compounds comprising amino acids linked by peptide bonds. These compounds may be modified, combined, complexed, derivatized, and so on.

Edman reagent or Edman type reagent means any reagent that is able to couple to all types of amino acids, and to sequentially bind to and cleave the amino acids from a polypeptide to form an Edman derivative, the derivative of each amino acid being able to be resolved on separation as a distinctive peak. The reagent must be able to do this with all amino acids at high efficiency (e.g. greater than 90%, preferably greater than 99%), and with essentially no side products that would interfere with resolution and detection ofthe peak. Sequencing Of Peptides And Proteins

Phenyl isothiocyanate (PITC) degradation (Edman, 1950) is the most efficient method for the sequencing of both small peptides and large proteins. The basics of this method are illustrated in Figure 1. There have been many modifications to the basic method, such as the liquid phase, or spinning cup sequencer (Edman and Begg, 1967), and an automated solid phase sequencer has also been described (Laursen, 1971). Both of these machines answered some of the limitations of the manual method. A gas-phase sequencer (Hunkapiller and Hood, 1978) is the current method of choice for sequencing both peptides and proteins. However, a rapid manual method is still used in some laboratories for chemical characterization of peptides and proteins isolated and purified chromatographically.

A comparison of manual, automated liquid phase, automated solid-phase, including the subtractive Edman sequencing with DABITC (4,4-dimethylaminoazobenzene 4' isothiocyanate), the dansyl Edman, and sequencing with other isothiocyanates have been reviewed by Allen (1981).

Amino Acid Analysis

An essential aid in the sequence analysis of peptides and proteins is automated amino acid analysis (Spackman et al. (1958)). The increase in sensitivity in microsequencing has been paralleled by a similar increase in sensitivity in amino acid analysis mostly due to four modifications (described below) in amino acid analysis methodology (Wilson, 1986):

(1) a change from ion exchange to HPLC (also resulting in a greatly reduced analysis time),

(2) use of microbore HPLC columns,

(3) precolumn derivatization, and

(4) the use of automation in vapor phase hydrolysis and derivatization of amino acids prior to amino acid analysis.

The net result is an amino acid analysis sensitivity of 1-5 picomoles. Of the 21 amino acids found in proteins, some are destroyed partially or totally during hydrolysis. The amounts are approximately 5-20% for serine, threonine, and tyrosine and 50-100% for tryp tophan, cysteine, and cystine. Antioxidants minimize these losses significantly. A recent modification of cysteine has been developed to yield a more acid- stable derivative. Pyridylethylation, using 2- or -4-vinylpyridine, produces a derivative that is easily quantified by both amino acid analysis and the Edman degradation and is eluted at a position unoccupied by other amino acid derivatives.

Chromatography The MPD™ system can quantify amounts of compounds as low as a hundred molecules, i.e. at sub-zeptomole (<10^~21 mole/sample) levels. See U.S. 5,866,907 and U.S. 5,854,044, incorporated herein by reference. By coupling HPLC with MPD readout (HPLC/MPD), these inventions provide a sensitive analytical technique which is applicable to separation, detection and quantitation of most analytes. For important biomolecules (neurotransmitters, steroids, insulin) limits of detection are about 1000-fold better than prior-art techniques. Sensitivity of 0.1 attomole/sample (10^"19 mole) sensitivity has been demonstrated.

Limits of detection (LODs) using HPLC/MPD: Chromatographic studies depend on the measurement of analytes which are often present at very low levels and current detectors are not sensitive enough for many applications. Even such sensitive methods as electrochemical detection and mass spectroscopy are often not sufficiently sensitive. Currently, the most sensitive chromatographic detection method compatible with Edman reagent at the level of a few hundreds of attomoles is laser-induced fluorescence (LIF).

Steroid analysis often calls for ultra-sensitive detectors. BioTraces, Inc. developed an ultrasensitive method for I labeled testosterone and estradiol. The separation of iodinated steroid hormones (Amersham) was performed using a Perkin Elmer instrument with Binary LC Pump 250, PE Pre Column Scavenger and C₁₈ Cartridge Column. To measure the sensitivity of HPLC/MPD, the individual compounds were dissolved in a methanol/water (9:1) mixture and dilutions from 100 pg/ml to about 100 fg/ml were prepared. Samples were 10 microliters, i.e. 1 fg/sample for the highest dilution. These solutions were submitted to HPLC and the fractions were collected and measured using the MPD device showing radioactivity of between 100 pCi and 0.01 pCi. For each steroid, the column eluent contained only one peak and its intensity was very consistent with the original sample radioactivity. In follow up experiments, radio-iodinated testosterone and estradiol derivatives were mixed, submitted to chromatographic separation under the same conditions as in the previous experiments, collected in 1 ml fractions and measured using the MPD device. Even for mixtures of steroids containing less than 1 fg/sample, one can see very well resolved peaks. These mixtures were the result of serial binary dilutions of the first mixture. Figure 2 shows the excellent linearity of the measurements of binary dilutions. The last dilution corresponds to a concentration below 10^" mole/sample. Even for this ultra-low concentration, which was undetectable in prior-art systems, the signal can be distinguished easily from the background (S/B > 50). In studies of inter- and intra- variability of HPLC/MPD, even at previously undetected levels, HPLC/MPD is very reproducible (see Figure 3).

HPLC/MPD excels in analysis of peptides and proteins. Attomole sensitivity has been demonstrated for a radio-iodinated peptide-insulin (6,000 Daltons). A solution of iodinated porcine insulin (7 x 10^"16 mole) was prepared by diluting a commercial ¹²⁵I labeled insulin (DuPont) in water and acetonitrile (50:50). Replicants were prepared by diluting commercial I labeled insulin in 50:50 wateπacetonitrile or urine (human female): acetonitrile in buffer and urine before distribution as 100 1 aliquots. Serial dilutions at a factor 1/4, starting at 1.75 x 10^"16 mole down to sub-attomole were used. Reverse phase HPLC analyses were preformed on 100 1 samples, injected manually, at ambient temperature on an Alltech Kromasil C8 column (25 x 0.46 cm, particle size 5 m ) with a guard column of the same packing. Eluent A was 0.1% trifluoroacetic acid (TFA) in water and eluent B was 0.1% TFA in acetonitrile. The solvent gradient used was 20% B increasing linearly at a rate of 2.5% a minute to a final concentration of 60% (held isocratic for 10 minutes). The flow rate was 1.0 ml/min and fractions were collected every minute for 20 minutes for counting by MPD. Figure 4 demonstrates a very well separated insulin peak for attomole/sample concentration of insulin, with the background below attomole/sample. The linearity of HPLC/MPD is shown in Figure 5. The reproducibility for dilute sample (a few attomoles/sample) is shown in Figure 6. As demonstrated above, the limits of detection when using HPLC/MPD for appropriately radioiodinated compounds are more than 1,000 -fold better than prior-art chromatographic methods.

The techniques discussed above can also be applied to other separation methods such as gel electrophoresis, capillary electrophoresis and thin layer chromatography (TLC). Applying MPD technology to measurements of samples separated by classical electrophoresis gives results very similar to those obtained when using HPLC. MPD detectors are also applicable to capillary electrophoresis. Additionally, the MPD detector can be coupled with thin layer chromatography (TLC). TLC/MPD techniques use the MPD Imager, which can quantify a TLC plate in few minutes for samples at the level of tens of attomoles.

According to the invention, an increase in sensitivity for polypeptide sequencing is achieved by:

1. Advanced technology for microsequencing and sensitive identification of the released amino acid derivatives

2. Powerful methods and radioiodination compatible with HPLC/MPD and TLC/MPD 3. Sophisticated isolation and purification ofthe protein or peptide sample under conditions suitable for microsequence analysis thereafter 4. Very high purity grade of reagent and solvents to allow for high quality sample purification and degradation

The invention permits peptide/protein sequencing at better than 100 attomole level, i.e. more than 1000-fold improvement over prior art methods, and to sub-attomole level.

General Approaches to Improved Sensitivity:

The main limitation of prior-art microsequencing is its sensitivity. Typically, a few

1 ") picomole (10^" mole) of peptide or protein is necessary to achieve reliable microsequencing. This leads to a very serious limitation on analytical and biological applications. Technically, the limitation on the sensitivity in all prior-art methods is due to the following: * limited sensitivity of prior-art chromatographic detectors used to resolve the chromatogram;

* problems of substance purification, i.e. the appearance of spurious peaks due to NSBB;

* problems of differential mobility due to use of "bulky" derivatization reagents;

* problems of compounds loses (NSL), especially of oxidation and hydrolysis for the case of some of amino acids such as cysteine or serine.

For example only two reagents used for precolumn derivatization have found extensive application. PITC yields a product with high extinction coefficients and has become the reagent of choice because PITC-amino-acid derivatives are relatively stable, with up to 60- min incubation in an aqueous buffer at room temperature. Another reagent is OP A, which produces a highly fluorescent adduct but OPA-2-mercaptoethanol isoindole derivatives of lysine degrade up to 30- 40%, histidine up to 20%, and many others up to 10% for the same time period. When thiols are used, e.g., ethanethiol or 3-mercapto-l-propanol, degradation rates are partially reduced. A second disadvantage is the failure of OPA to react with secondary amines, proline and hydroxyproline. For this reaction to occur Na hypochlorite is used as an oxidant. The detection of these amino acids following this oxidation step has found only limited application in post-column detection systems and the quenching characteristic of Na hypochlorite and its oxidation of other amino acids limits sensitivity and makes the OPA system less desirable. When losses with OPA are compared over a 12-h period, decreases range between 2 and 50%. In those cases where apparent increases were noted, i.e., glycine and proline, co-elution of breakdown products with the derivatized amino acids was found responsible for these increases.

Furthermore, there are technical limitations on the sensitivity and reliability of microsequencing methods using fluorescein conjugated to Edman reagent:

1. Degradation of polypeptides with FITC is less suitable for the liquid phase method. Due to fewer contaminants, sequencing in the solid phase manner was more efficient than by the liquid phase manual method. 2. Comparison of the FITC reagent with DABITC showed no increased sensitivity: the higher detection sensitivity of the fluorescent isothiocyanate is compensated by the lower coupling yield of this reagent, probably caused by steric hindrance ofthe bulky chromophore. Especially the N-terminal residues react poorly, which may be caused by repulsion ofthe negative groups between the dye and the peptide.

3. FITC may decompose during the degradation. The reagent carries many functional side groups that can give rise to many side reactions. Consequently, many extra spots are visible on the thin-layer sheets which obscure the degradation result. Therefore, higher quantities of the peptide must be subjected to the sequence analysis than with DABITC or PITC.

4. Removal of the residual FITC reagent after the reaction is more difficult, even with extensive wash cycles.

5. Separation of all PTH-amino acids by two-dimensional thin-layer chromatography on polyamide sheets is not possible: PTH-Met/Phe/Nal and PTH-Ile/Leu are poorly resolved.

6. Non-covalent adsorption of the FITC fluorochrome to proteins cannot be neglected. Sequencing of proteins is hampered by this effect. Further, adsorption of FITC to the glass support seems possible.

The second problem can be largely diminished when using more sensitive detection means. Each purification method leads to a certain, often large, loss of studied analyte. With more sensitive means of detection, more specific methods of purifications can be used, because one is not limited by the need to extract almost all present material. The problem of differential mobility is drastic when using pre-column derivatization, e.g. when using fluorophores as labels. The molecule modification by an attachment of a single atom of radioiodine is much lower, and the differential mobility either negligible or can be calibrated. In the disclosed implementation using radioiodinated Edman reagent, the "two color" labeling available with ¹²⁵I and ¹²³I can be used to diminish the uncertainties due to nonspecific losses. For example, relative measurement methods with use of ¹²³I for standard and ¹²⁵I label for unknown are very efficient in the diminishing NSL, i.e. leads to much higher reliability of peptides and proteins microsequencing. The most commonly encountered problem in protein sequencing is modification of the N- terminus. Existing methods for deprotection are in many cases not efficient enough (and cannot be improved without a risk of protein degradation) to provide sufficient amounts of free amine group available for the first cycle. In such cases improved sensitivity would allow sequencing of minute amounts of deprotected proteins.

In many applications isolation of the protein is achieved by blotting appropriate gel on the membrane. Modern gas-phase sequenators allow direct use of the blot for sequencing. However, a common problem is insufficient amount of the protein transferred to the membrane. In this case multiple blotting or repeated electrophoresis might be necessary. This leads to multiple layers of blotting membrane in the chamber, and as a result less reliable reaction yields. Use of much smaller amounts of the proteins or mixture thereof significantly improves resolution ofthe gel, and thus purity ofthe final blotted sample.

Finally many analytical protocols both in industrial as well as research applications require characterization of the protein products by peptide mapping, followed by sequencing of amino acid analysis. Separation of frequently complicated mixture of peptides is greatly improved when the scale is diminished. Subsequent analysis of small amounts of pure peptide can be significantly facilitated by MPD™-enhanced techniques.

Importance of increased sensitivity in protein sequencing:

Increased sensitivity of the available protocols used to sequence protein is needed in many cases. Indeed, the technologies involved in protein purification have made remarkable progress in recent years and it is now quite feasible to purify proteins in femtomole amounts. Further characterization or identification of these purified proteins is at present impossible when their amount is below a few picomoles. Sensitive sequencing according to the invention enables many new methods.

1) Protein sequencing made easier and more reliable.

First, some proteins that can already be purified in picomole amount can be more easily and reliably sequenced: a) a longer stretch ofthe sequence can be achieved: the yield of each step is such that the amount of cleaved amino acid decreases in an exponential way as the sequencing proceeds. Therefore, the Edman degradation has to be stopped after a certain number of cycles, depending on the amount of starting material. Sometimes, the number of cycles performed is not enough to allow unambiguous and/or rapid identification ofthe protein (e.g. impossibility to identify variants ofthe protein such as splice variants, post-transcriptionally modified variants, or design of too small or too degenerate oligonucleotides to screen a cDNA library). In some instances, the complete sequence could even be achieved, and the need of still difficult C-terminus sequencing would become dispensable.

b) smaller quantities of sample are required:

- proteins present in smaller samples could still be reliably identified. For instance, identification of proteins present in precious samples like tissue biopsies or biological fluids could be carried out from smaller samples.

- requirement of smaller samples could also translate into scaling down of the upstream steps of the process leading to the purified protein. For instance, smaller sample could mean less starting material and less reagents to prepare this material.

- smaller samples could also mean higher density of different samples processed in parallel experiments, as is frequently the case in some drug-discovery protocols in which a large number of proteins are screened in parallel. Since the sequence of only a small subset of these proteins need to be eventually determined, protein sequencing output is not the rate- limiting factor. But poor sensitivity of the sequencing translates into larger samples and therefore, fewer samples that can be screened in parallel.

2) Sequencing of a new set of proteins available in sub-picomole amounts can be performed.

The invention permits study of proteins poorly expressed in cells or tissues like transcription factors or variants of proteins that are poorly represented in a tissue (e.g. mutated oncogene in a tumor sample, in which the majority of the sample protein is unmutated). The level of global protein expression, termed "proteome", can be assayed by means of two-dimensional gel electrophoresis. Typically, 10 micrograms of protein can be loaded on such gels, which can be resolved into as many as 5000 spots of higher intensity. But many important proteins are present in low or undetectable levels by conventional methods. The MPD-enhanced detection of this set of poorly expressed proteins has now become possible, as well as their purification from 2D-gels. MPD-enhanced Edman degradation then allows the identification and sequencing of these proteins, which has remained an unreachable goal for the man ofthe art using existing methodologies.

Another application that benefits from MPD-enhanced Edman degradation is the sequencing of proteins with a "blocked" N-terminus. A significant fraction of proteins undergo post- translation modification resulting in N-terminus blocking, which prevents the coupling ofthe Edman reagent. Various protocols and reagents have been developed aimed at unblocking this N-terminus, but these usually achieve low or unpredictable yields (e.g. anhydrous hydrazine vapor; yield can be in the range of a few percents (Eur. J. Biochem, 1993, 212:785), or heptafluorobutyric acid (HFBA) vapors (Electrophoresis (1996), 17:855)).

Frequently, the yield of the unblocking reaction is so low that the unblocked protein is in insufficient amount to be sequenced. An alternative is to partially digest the N-terminus blocked proteins and sequence some of the peptidic fragments upon purification. However, this is time consuming and the N-terminal sequence of the protein is still unknown. MPD- enhanced Edman degradation solves this issue by allowing the sequence determination of a large proportion of these N-terminus-blocked proteins, upon unblocking with already existing methods and reagents despite their low yields.

3) Further applications of supersensitive methods of peptides/proteins sequencing:

The primary structure of proteins or peptides can be elucidated in two ways: a direct method consists in sequencing the protein using sequencing methods, the most widely used being the Edman degradation. The second, and indirect method, consists in determining the sequence ofthe gene or its transcript encoding this protein, with the assumption that this gene has been identified and that its sequence has been previously elucidated. A conceptual translation based on the knowledge ofthe genetic code provides the sequence ofthe protein. However, the actual sequence of the protein may differ from the sequence derived by conceptual translation for a variety of reasons: first the genetic code is not universal; some organisms or organelles (e.g. mithochondrium) use variants of the code used by eukaryotes; the genetic code may even be altered in a given cell under some circumstances (e.g. stop codon suppressor phenotype in bacteriophage-infected bacteria). Second, the exact sequence of the transcripts of a given gene is not fully predictable: in eukaryotes, they may be differentially spliced according to tissues or time of expression, or even edited in some occasions, so that the mere knowledge of the gene sequence is often insufficient to deduce the protein sequence. Third, even if the transcript sequence is used for conceptual translation, its translation may involve some so far unpredictable alterations such as cryptic initiation of translation, frameshifting or translation by-pass. Last but not least, proteins can undergo, after translation, some of many possible post-translational modifications of their primary structure that may be either definitive or time-dependant (e.g. protein cleavage, myristilation, deamidation, glycosylation, phosphorylation).

Radio-derivatization compatible with MPD enhanced HPLC:

Pre-column derivatization is an important method in peptide/protein sequencing. By analogy to other derivatizing agents used in chromatography, such as luminophores and fluorophores, we call these agents radiophores.

There are two options for radiolabeling that are compatible with Edman degradation process.

The first process takes advantage of the presence of an activated aromatic ring in some natural amino acids (tyrosine and histidine) and relies on electrophilic aromatic substitution with iodide ion in the presence of an appropriate oxidizing agent such as chloramine T. This method permits the radioiodination of two out of 21 amino acids. It is, very important however, because it permits one to calibrate the losses that occur during the process, e.g. tyrosines and histidine can be radioiodinated with ¹²³I, and total activity measured before Edman degradation process. Measurement, after each step of process, permits reliable estimate ofthe step yield. The main novelty of the disclosed method is to radiolabel, preferably radioiodinate the Edman reagent which during the microsequencing of peptides/proteins is reacted with N- terminal amino acid. There are three major reactions used to introduce atoms of iodine- 125 into the aromatic ring:

* Sandmeyer reaction of diazonium salts synthesized from aromatic amines;

* nucleophilic substitution of aromatic bromides and iodides;

* electrophilic aromatic substitution using [¹²⁵I]iodide in the presence of an oxidizing agent.

Electrophilic aromatic substitution is a very effective method. Unfortunately, it requires that the iodide ion is oxidized to iodine atom. Two atoms form a molecule of iodine which is the immediate source of the electrophile. Molecular iodine is highly volatile, and thus in the case of radio-isotope, requires special safety precautions. The aromatic ring of the iodinated compound preferably contains an electron donating group such as hydroxy, alkoxy or amino to facilitate an electrophilic reaction.

Radiolabelling of Edman-Type Reagents:

Pre-column Labeling: In a preferred method, where the least possible modification of the existing procedure is used, 4-[¹²⁵I]-iodophenyl isothiocyanate and ^I253-[I]-iodophenyl isothiocyanate can be synthesized and substituted for PITC. 4-[ I]-iodophenyl isothiocyanate may be obtained by isotope equilibration of 4-iodoaniline, and subsequent conversion of the amino group into isothiocyanate with carbon disulfide in the presence of dicyclohexyl carbodiimide (DCC)(Burrel et.al. 1975). The resulting 4-[ I]- iodophenyl isothiocyanate was used for Edman degradation cycle with separation of 19 [¹²⁵I]-iodo- PTH- amino acids by means of silica gel TLC. However, the described procedure was performed only in three cycles, and the sensitivity of the method was poor, in the range of few nanomoles. Apparently the yield of iodinated product was low, and impurities were present.

The reagent was therefore inappropriate for microsequencing as evidenced by its disuse for over two decades. Higher purity and yields are accomplished according to the invention.

3-[ I] -iodophenyl isothiocyanate may be obtained in one step by electrophilic substitution of the tri-«-butylstannyl group in meta position of the PITC with I in the presence of oxidative reagent (Iodo-Gen)(Ram et. al. 1994). 3-[¹²⁵I]-iodophenyl isothiocyanate in this report was used for antibody labeling without any reference to protein sequencing. In the present invention we disclose alternative synthesis of [ I]-iodophenyl isothiocyanate from iodoaniline and thiophosgene (see Fig. 7).

It is understood that this preferred implementation is an example of the simplest possible modifications leading to radioiodinated aromatic isothiocyanate. Derivatization procedures may include activating substituents in the phenyl ring such as 4-hydroxy or more preferably 4-methoxy. Equivalent multi-ring aromatic systems can also be used in the disclosed method such as naphthalene, anthracene, phenanthrene, pyrene, benzidine, biphenyl, and the like. It is also understood that any combination of heteroatoms can be present in the aromatic ring and the examples of such systems include: N-alkyl pyrane, thiophene, pyrimidine, pyridine, indol, phenantrolines, and the like, where arrangement of the iodo and isothiocyanyl substituents, possibly together with other substituents, would result from rational chemical design used in the chemistry of aromatics, and taking into consideration availability of the substrates, inductive effects, efficiency ofthe chemical steps, and the cost ofthe reagents.

Possible labeling methods are not limited to the ring of the isothiocyanate. In an alternative derivatization method cyclized 2-anilino-5-thiazolinone (ATZ) amino acid is reacted with primary amino group of labeling reagent e.g. [¹²⁵I]-iodohistamine or 4-aminofluorescein (Tsugita et.al. 1988, and 1989). This reaction allows the use of a plethora of new reagents capable of being labeled.

Post-Column Labeling: Another alternative Edman reagent was developed using [4- (tert- butoxycarbonyl-aminomethyl)]-phenyl isothiocyanate (t-BOC-aminomethyl-phenyl isothiocyanate, BAMPITC). BOC is an acid labile protecting group that is removed during cleavage step in the presence of trifluoroacetic acid. Deprotected 4-aminomethyl group is unchanged during HPLC analysis. It has been shown that despite increased hydrophilicity which decreases retention time, resolution was achieved for all amino acids. Again this example shows the simplest possible method to introduce primary aliphatic amine function into the Edman reagent. This modification is not the only one that is explored in the ongoing research. Similarly, the amine protecting group can be modified, and the use of monomethoxytrityl (4-methoxy-triphenyl methyl, MMT) is considered. An alternative for the above protocol is the use of B AMPITC in the sequenator, and labeling prior to HPLC. This procedure involves an additional step of 4-aminomethyl reaction preferably with Bolton-Hunter reagent (radioiodinated derivative, see Figure 8), and subsequent HPLC analysis.

A radiolabeled equivalent of Bolton-Hunter reagent and radioiodinated pipsyl chloride have been shown to be effective for specific labeling ofthe amino groups post-separation in MPD- enhanced chromatographic techniques.

Optimization of Radioiodination of Edman Reagent: All reactions leading to labeled PTH- amino acids are preferably optimized using "cold", non-radioactive, iodine derivatives. For example, 4-iodoaniline is commercially available and is converted into corresponding 4- iodo-PITC with thiophosgene. Optimization of this process comprises simple purification of the product without gel chromatography. Similarly, reaction of B AM-PTH-amino acid with Bolton-Hunter reagent is preferably optimized using anhydrous conditions for the final yield, and removal of excess Bolton-Hunter reagent. After hydrolysis, it is extracted from highly lipophilic product with a buffer. Alternatively, micro gel filtration on normal phase, using lipophilic eluents, may be used.

A third method is the reaction of 2-iniline-5-thiazolinone (ATZ) amino acid with labeled primary alifatic amines, preferably attached to heterocyclic aromatic ring. In the most preferable implementation, [ I]-iodohistamine is used. Chromatography under control of MPD instrumentation can be used to minimize impurities to the attomole level.

Influence of ¹²⁵I label on products of Edman degradation: Analysis of the effect of conjugation of iodine on the reagents ring showed that these changes are negligible. For example, minor changes of induction effect in the phenyl ring of PITC was shown not to influence the kinetics of the phenylcarbamoyl moiety formation. Subsequent steps are more influenced by the side chains of amino acids than substituents of the aromatic ring. The important consideration is final separation of 4-[¹²⁵I]-iodo-PTH (or alternatively, [3-(4'- methoxy-3'-[¹²⁵I]- iodophenylpropionyl]-4-aminomethyl-PTH) derivatives of amino acids. Criteria for Reagents:

The criteria for selecting a promising reagent are:

1. The reagent should be available following simple synthesis routes which allow it to be purified easily to the highest possible grade. 2. The reagent should be volatile. This permits the removal of the excess reagent after the reaction without losses of peptides or proteins. PITC and methylisothiocyanate (MITC) are preferred in this regard. 3. The coupling of the reagent to the free amino groups of the proteins should exceed 90%; the reaction must be as complete for hydrophilic and small residues as for those with bulky and hydrophobic side chains. This is only possible if the reactive group of the reagent is not sterically hindered.

4. The reagent should provide a radiophore prior to or after the degradation, which enables sensitive detection ofthe released amino acid derivatives. If the radiophore is to be provided after degradation, the reagent should have a protected group for binding the radiophore which is deprotected after coupling to the amino acid.

5. Complete cleavage ofthe coupled residue from the polypeptide chain at the cleavage should not be sterically hindered by the radiophore part ofthe reagent.

Edman-type reagents according to the invention meet all these requirements. Many different isothiocyanate-type reagents are compatible with I radioderivatization. Only the classical PITC reagent has been studied in much detail, but synthesis of other suitable radiolabeled isothiocyanate homologues can be accomplished by persons of ordinary skill.

For example, radiodinated dansyl-PITC (DNSAPITC) is suitable for sequencing; the reagent produces PTH-amino acid derivatives which are detectable in low picomole quantities employing HPLC separation and detection with a fluorescence detector. In addition, the application os small-sized polyamide thin-layer sheets for additional identification of 1 to 10 pmoles has been demonstrated. This allows safe microsequencing of high reliability. The excellent stability of the derivativess add to the quality of this Edman-type reagent. Using radioiodinated reagent, manual sequencing employing DNSAPITC can be performed on the 50-100 attomole level and with many samples at the same time. Hence this reagent serves as a possible alternative to the classical PITC or DABITC/PITC sequencing approaches.

Use of other MPD compatible isotopes for derivation of Edman-Type reagent:

Among the MPD technology compatible isotopes the largest group (about 60) belong to the lanthanide group, and lanthanide complexes can be used in the chemistry of the Edman degradation. It has been shown that lanthanide complexes are sufficiently stable to be used for HPLC (Okabayashi et al., 1994). Chelating agents for lanthanides include N-benzyl diethylenetriaminetetraacetic acid. In a preferred implementation, the chelator is N-(p- isothiocyanatobenzyl) diethylenetriaminetetraacetic acid.

Eu⁺³ chelate fluorescence detection is known to have many limitations, especially in HPLC, where e.g. increase of acetonitrile concentration above 20% causes rapidquenching of fluorescence. Quenching introduces significant errors; thus fluorescence based mehtods are not quantitative. The innovative substitution of a radiolanthanide for fluorescent Eu comples avoids problems of fluorescence quenching which have been observed in prior art (Mukkala et al., 1989). Other chemically campatible chelates can be used, for example trisbipyridine cryptates (Lopez et al., 1993). MPD-enhanced detection is fully quantitative, as radioactivity is independent of chemical or physical factors.

Microsequencing Polypeptides Using Radioiodinated Edman-type Reagent:

The inventive use of radioiodinated Edman reagent is fully compatible with the overall chemistry of the Edman degradation in all current protocols. Modern microbore HPLC supports have superb separation efficiency. According to inventive methods, where various substituents are used in aromatic isothiocyanates, it is almost always possible to achieve full separation of the amino acids in HPLC. Choice of the MPD-enhanced methodology is also based on separation feasibility. Introduction of iodine in the phenyl ring does not change lipophilicity ofthe PTH-amino acid, and thus provides very minute variation ofthe retention time. Based on the HPLC evaluation on non-radioactive models, High-Performance thin layer chromatography (HP-TLC) and Reversed Phase TLC (RP-TLC) may be used to diminish the cost of instrumentation significantly, as compared with HPLC or capillary electrophoresis with Laser Induced Fluorescence detection (CE-LIF).

An important feature of the present invention is its ability to provide a highly parallel sequencing process. In modern molecular biology, there is an increasing need for high- throughput analysis of proteins. In the discovery process, it is important to analyze only part ofthe protein in order to design a hybridization probe for gene library screening. Many very important proteins are present in minute amounts, because they provide very powerful signals for the transduction cascades. Thus, sensitivity of protein detection, and especially super-sensitive, high-throughput sequencing can significantly facilitate protein analysis. Embodiment of the invention can be used for parallel sequencing of many (ca. 50) proteins from 2D gels at the zeptomole level, where Edman degradation products can be separated and analyzed (detected and quantitated) simultaneously on a TLC plate. TLC plates are compatible with the MPD-imager format, where detection of the peaks is very rapid, and then it is possible to quantitate the peaks with increasing and adjustable spatial resolution. This approach has lower cost than CE-LIF or especially HPLC/MS.

The efficiency of the chemical reactions is important. Because the amount of analyzed sample is so small, the reaction rate at the coupling step is diminished according to kinetics. It is therefore necessary to use the reagent in excess. Other steps (cleavage and conversion) are intramolecular and their rate is concentration independent. Acid catalysis, especially gas- phase, provides stoichiometric excess of the TFA, so appropriate protonation steps are not rate limiting.

TLC is an advantageous separation medium for analysis of the Edman degradation products. Although this method was used with success in 1960s, the was a complete shift to HPLC and CE in automated sequenator design. However, use of High-Performance or Reversed Phase gels in TLC allows at least equal separation efficiency at lower cost and much higher throughput, and TLC readout is preferable for multi-photon detection than HPLC, because of the multi-photon detection-imager characteristics.

Protein microsequencing may also be carried out with a modified HPLC readout, using MPD techniques. In classical implementations, HPLC is used to fractionate the amino acids in "flow" mode. Thus, an appropriate optical detector (UV absorbtion, fluorescence) is placed at the end ofthe column and the amount of reagent derivatized to increase the optical signal, either emission or absorbtion, is measured when it passes through the small size optical element. This implementation is not efficient when using ¹²⁵I as a label, because the time of transit through the active zone is very short, typically about 1% of total time of separation. However, the radioactive decay process is time-extensive, i.e. statistically the probability of decay is constant over the time of measurement if it is much shorter than the half-life of the given isotope. It is possible to overcome this limitation of HPLC-MPD technique read-out with a variety of scanning and imaging techniques based on the knowledge that only one I labeled peak is expected to appear as the separation output.

In one embodiment, the time of transit of a given Edman degradation product through a 20 cm long HPLC column is 20 minutes and the fraction has a concentration of about 50 attomole, corresponding to 100 decays per minute (100 dpm). The separation ofthe Edman degradation product happens within the first part of the HPLC column, even if the optimal spatial resolution (fully formed, narrow peak) is obtained only at the column exit. The full width, half maximum (FWHM) of this peak is 1 cm at the distance of 10 cm and 0.5 cm at the column exit. The peak displacement speed can be reliably calculated from its movement in the first part of the column. A one dimensional MPD spacially resolving detector with aperture of about 2 cm and spatial resolution of 0.2 mm (SR-MPD) is placed at about the halfway point of the HPLC column. When ¹²⁵I labeled product enters the SR-MPD detector's aperture, the count rate is sent to a computer which calculates the peak profile. The peak can be established within about 20 sec with a precision of about 5 mm. Also, the peak displacement speed can be calculated with better than 10% precision. The SR-MPD device can be co-moved with the peak and during the next ten minutes a full profile of activity in the peak is acquired. For 50 attomole of Edman degradation product, the total amount of counts acquired is calculated to be about 1,000 which permits quantitation of amount of analyte with statistical precision better than 3%. Also, the peak position can be established with precision better than 1 mm.

Because MPD detectors are virtually zero background devices, qualitative detection of minute amounts, e.g. only three counts, is statistically meaningful. That is, when using MPD one can very quickly establish that a given "pixel" or "voxel" contains I or another MPD- compatible label. However, quantitation may require a longer time. Typically one would like to acquire at least 100 counts/pixel to obtain 10% statistical uncertainty level.

In another embodiment, the diverse fractions of HPLC output are retained on the surface of appropriate filtering medium, e.g. a moving band of filter paper. In this application, the length ofthe chromatogram is between half and twice the length ofthe HPLC column. First, all of the chromatogram is very quickly scanned by an MPD device for the presence of a peak. Next, a high precision but much slower quantitation process is performed.

This quantitation method is compatible with the MPD scanner. Accordingly, a MPD detector is equipped with changeable lead aperture and the object to be scanned, e.g. a one- dimensional chromatogram, is mechanically moved past the aperture. By way of example, the full length ofthe chromatogram is 20 cm, i.e. about the same as the length of typical high resolution HPLC column. The width of chromatogram is arbitrary, but in a preferred implementation it is smaller than 1 cm. The MPD detector may have a diameter of 2.54 cm. The first step is to scan the chromatogram with the "wide aperture" with dimensions of 2 cm x 1 cm. Thus in ten steps the whole length of the chromatogram can be scanned. If the activity of the Edman degradation product is 50 attomole, within about 10 seconds per position one can estimate reliably if the peak is within the aperture. This screening process is very reliable: both "false positive (FP) and "false negatives" (FN) are controlled to be smaller than 1%. In the next step a "medium aperture" of 5 mm x 1 cm is introduced. A scanning process with the spatial displacement value of 0.2 cm is used with an acquisition time of about 20 seconds. Within less than 3 minutes, i.e. 10 steps, the position of the peak is established with a precision better than 2 mm. Once more the aperture is reduced in size to 1 mm x 1 cm, and scanning with steps of 0.2 mm is initiated, each step taking 30 seconds. Thus, within less than 5 minutes, the position of the peak is established with a precision better than 0.5 mm. Additionally, the activity profile of the peak is established with good precision, i.e. the number of counts under the peak is established with a precision better than 5%. The surprising implication is that, in the case of Edman degradation analysis, even a simple MPD scanner permits one to establish the peak position and amplitude in less than 10 minutes, i.e. in a time shorter than the complete time of separation by the HPLC. An added advantage of this method is that it allows the chromatography process to be stopped before any Edman degradation product leaves the column. In an embodiment to accomplish this goal, a standard with retention time slightly lower than first Edman degradation product is loaded at a level of at least 100 femtomole into the column. HPLC apparatus with an in-flow optical detector may be used to detect the transit of the standard marker. Then, the pressure is released leading to a "frozen" pattern of the Edman degradation products inside the column. Such a column can then be scanned or imaged by an appropriate MPD detector. This is feasible because high energy X/gamma rays emitted by I easily cross the thin wells of the column, even if stainless-steel columns are used. The above-described strategy of repeated interactive MPD scans with three variable dimensions aperture can also be implemented in this case. However as the column typically has less than 2 mm diameter as compared with a 1 cm chromatogram as described above, appropriate apertures are 2 cm x 0.5 cm, 0.5 x 0.5 cm and 0.2cm x 0.5 cm, respectively. The quantitation process will typically take from a few minutes to about 30 minutes for less than 20 attomole fractions. In this time, diffusion may lead to considerable smearing ofthe peak. Thus when measurement time is expected to be above a few minutes, the diffusion should be diminished by cooling the column, preferably below the freezing point ofthe buffer, or other means.

In a preferred implementation, the multicolumn HPLC system is used and the "freezed" pattern is obtained essentially simultaneously in all columns in parallel. Thus, the columns can not be mechanically removed and placed onto an appropriate 2D MPD Imager. In one implementation using multiple HPLC columns and an SR MPD imager (4 columns x 10 detectors per column) is possible using a 2 inch SR MPD. This arrangement leads to about 20 fold faster read-out than when using a single detector. Multiple crystals read by a photodiode array can achieve a further 10 fold acceleration ofthe quantitation process.

Amino Acid Analysis Using Radioiodinated Edman-Type Reagent:

In another embodiment, quantitative amino acid composition analysis may readily be achieved without obtaining sequence information. The analysis involves three steps:

* acid catalysed exhaustive hydrolysis ofthe protein; * Edman reagent conjugation step, followed by cyclization and conversion;

* chromatographic separation of all PTH-amino acids.

In modern amino acid analysers, acid hydrolysis of the protein is achieved using gaseous hydrogen chloride at elevated temperature. Subsequent conjugation, cyclization and conversion steps are performed in conditions identical to those used in sequencing. The resulting mixture of PTH-amino acids is injected on the HPLC column.

A supersensitive MPD-enhanced quantitation process according to the invention starts with

19^ detection of I in PTH-amino acids mixture before starting the separation process. This permits fast, statistically meaningful estimation of the total amount of Edman degradation product and facilitates "computer" based optimization of the quantitation process. When quantitating the outputs of HPLC for an arbitary analyte, neither the number of peaks nor their relative amplitudes are known. The total number of peaks can be large, say up to 100 and often two peaks for different analytes are partially overlapping. Also the dynamic range is often large, say two logs. When studying the outputs of Edman degradation process, a considerable amount of information is a priori available. First, the total number of peaks is no greater than 21 and often is lower. Second, the peak are well resolved and their position is well known a priori. Third, the dynamic range of peaks amplitude is limited, typically less than 1 log. Actually, when the products of each cycle are fractioned separately there is only one peak and for all cycles the amplitude is the same. When the outputs of different cycles are pooled, the amplitudes of different peaks after normalization to the highest peak generally give a characteristic pattern of fractional numbers for amplitudes, i.e. amplitudes are 1, 1/2, 1/3, 1/4...1/i wherein "i" is integer number. Finally, generally if the protein itself has been purified by HPLC its length is known. Using this information permits considerable acceleration of the process and improvement of reliability, when quantifying the HPLC output.

In a preferred implementation, the end part of the HPLC column is shaped with a constriction ofthe appropriate size. Thus, according to Bernoulli's law, the flowing liquid is accelerated and can be induced to break into droplets of very small, submillimeter diameter. In another implementation, the process of creating drops is accelerated by applying to the end part of the HPLC column vibrations of appropriate amplitude and frequency. Optionally, high frequency pressure can be applied through the HPLC itself, i.e. the constant or gradient pressure operation of HPLC is modulated by a high frequency but relatively low pressure component. In yet another implementation, the end part of the HPLC capillary is made of an elastic material, e.g. plastic tubing whose geometrical dimensions are modulated by an external actuator. The said actuator can be either mechanical, electromagnetic or piezoelectric. In another innovative implementation, the HPLC column end is placed into a high pressure gas chamber in which the pressure is modulated leading to an oscillating differential pressure which induces droplet creation. Similarly, by fast cycling of the pressure in the "gas chamber" from vacuum to atmospheric pressure, the creation of droplets can be effected.

The selected HPLC conditions may involve the use of a pressure gradient leading to time variable flow speed. Thus, the diameter and number of created droplets may be time dependent. However, it is advantageous to have a constant diameter of droplets which can be achieved by establishing a feed-back between the pressure and the properties of a device responsible for breaking the flow into droplets.

The drops are collected on the moving band consisting of the absorbent material. The properties of the material, e.g. porosity, are selected so that all amino acids are retained but the HPLC buffer moves freely across the band of the moving material. When using porous material as absorbent of the droplets from HPLC, the process of diffusion inside the porous material may lead to loss of resolution. Thus, the absorbent material may be divided into millimeter size pixels of absorbent material characterized by high diffusivity, and submillimetric walls made of material with substantially lower diffusivity. Typically, the size of the pixels is 0.5 to 5 mm and the walls are less than 0.2 mm thick. The preferred implementation of such a pattern is by controlled spraying of nonporous plastic upon a filter paper band.

It is advantageous that statistically one drop is absorbed per each pixel of moving band. Thus, a chromatogram can be interpreted as a sequence of 0 or 1 bits, and subsequent quantitation is considerably accelerated, i.e. only the domains which absorbed the droplet are quantitated. This can be achieved by feed-back between the appropriate optical detector sensitive to droplet creation at the orifice of the HPLC and the mechanism displacing the band. The appropriate information can be stored in a computer to be subsequently used for optimizing the scheme of detection using an MPD Imager. In another implementation, the buffer of HPLC can be tinted with an appropriate colorant. Thus the droplet pattern on the

19S moving band can be read optically by the appropriate optical detector and the I content of only colored pixels is measured by an MPD Imager.

In high throughput applications it is preferable to miniaturize the chromatogram dimensions. Thus, the chromatogram may be essentially two-dimensional and the mechanical system consisting of a computer steered x-y mover with at least 100 microns precision to permit moving the absorbing material in a zig-zag pattern similar to TV raster pattern. This produces a 2D array of pixels. Optionally, the x-y mover is operated to obtain a spiral pattern of droplets adsorbed on the surface of a chromatogram. Using CCD-based MPD imagers it may be possible to quantitate concurrently all pixels of 128x128 a pixels chromatogram.

TLC based implementations: Two-dimensional TLC ofthe PTH-amino acid products has been used in protein sequencing. Modern silica gels allow achieving resolution capabilities comparable to HPLC systems. Two TLC gel types high performance (HP TLC), and reversed phase (RT TLC) are fully compatible with MPD techniques. Resolution, sensitivity and reproducibility of MPD-enhanced TLC has been previously shown. HP TLC can be developed in a gradient of pH, while RT TLC can use a gradient of organic solvent (typically acetonitrile).

An advantage ofthe proposed implementation is an increase in the throughput ofthe parallel sequencing reactions. Single TLC can easily accommodate 50 samples that can be the products of 50 simultaneously performed sequencing operations on individual blot pieces cut out of the two- dimensional gel. At the end of each cycle the products are separated on separate TLC plate and analyzed by MPD imager. The increase in throughput compensates more than enough for the manual handling necessary in the simplest prototype versions. Additionally, the cost of TLC-based analysis is very moderate in comparison with HPLC equipment. Automation brings further advantages. The embodiments illustrated and discussed in this specification are intended only to teach those skilled in the art the best way known to the inventors to make and use the invention. Nothing in this specification should be considered as limiting the scope of the present invention. Modifications and variations of the above-described embodiments of the invention are possible without departing from the invention, as appreciated by those skilled in the art in light of the above teachings. It is therefore to be understood that, within the scope of the claims and their equivalents, the invention may be practiced otherwise than as specifically described. All references cited herein are hereby incorporated by reference in their entirety to the same extent as if each were individually incorporated by reference.

REFERENCES:

Allen G (1981): Laboratory techniques in biochemistry and molecular biology: sequencing of proteins and peptides. In Work TS et al, (ed): pp 174-180. New York: Elsevier Science Publishers.

Boehnert M, Schlesinger DH (1979): Improved manual sequential analysis of peptides.

Anal Biochem 96:469473.

Brauer AW et al, (1975): The application of 0. I M Quadrol to the microsequence of proteins and the sequence of tryptic peptides. Biochemistry 14:3029-3035. Braunitzer G et al, (1970): Versuche zum volistandigen und automatischen. Abbau von

Peptiden nach der Quadrolmethod. Hoppe Seylers Z Physiol Chem 351:1589.

Bruckner CA. et al. (1994) , Column Liquid Chromatography: Equipment and

Instrumentation, Anal. Chem., 66, 1R,

Butler WT (1969): The identity of a hydroxylamine-sensitive bond in the I chain of rat skin cobagen. J Biol Chem 244:3415-3417.

Capony JP, Demaille JG (1983): A rapid microdetermination of phosphoserine, phosphothreortine, and phosphothyrosine in proteins by automatic cation exchange on a conventional amino acid analyzer. Anal Biochem 128:206-212.

Church W.H., Justice J.B., Jr. (1989) Adv. in Chromatography, 28,165 Dorsey J.G. et al, (1994) LC: Theory and Methodology, Anal. Chem., 66, 500R

Edman P (1950) A method for determination of the amino acid sequence in peptides. Acta

Chem Scand 4:283-293.

Edman P, Begg G (1967): A protein sequenator. Eur J Biochem 1 :80-90.

Edman P (1975): In Needleman SB (ed): "Protein Sequence Determination." Berlin, Heidelberg, New York: Springer- Verlag pp 211 -279.

Fekkes D. State-of-the-art of high-performance liquid chromatographic analysis of amino acids in physiological samples. Journal of Chromatography B: Biomedical Applications

682(1). 1996. 3-22.

Fielden P.R., (1992) Recent Developments in LC Detector Technology, J. Chrom. Science, 30, 45

Goto M et al, (1989) Adv. in Chromatography, 30, 167 . Hunkapiller MW, Hood LE (1978): Direct microsequence analysis of polypeptides using an improved sequenator, a non-protein carrier (Polybrene) and high pressure liquid chromatography. Biochemistry 17:2124.

Hunkapiller RM, Hood LE, Dreyer WJ (1981): A gas liquid solid phase peptide and protein sequenator. J Biol Chem 256:7990.

Kemp BE (1980): Relative alkali stability of some peptide o-phosphoserine and o- phosphodireonine esters. FEBS Lett 110:308-312.

Krull I.S., Szulc M.E., Bourque A.J., Zhou F.X., Yu J., Strong R., (1994): Solid-phase derivatization reactions for biomedical liquid chromatography. J. Chromatogr., B, 659, 19- 50.

Laursen RA (1966): A solid phase Edman degradation. J Am Chem Soc 88:5344-5346.

Laursen RA (1971): Solid phase Edman degradation. An automated peptide sequencer. Cur

J Biochem 20:89-102.

Laursen RA (1975): "Solid Phase Methods in Protein Microsequencing of Peptides and Proteins 43

Lee HM, Riordan JF (1978): Solid phase Edman degradation: A new method for attaching peptides to insoluble resins using trifluoracetic anhydride. Anal Biochem 89:136-142.

Liu TY, Chang YH (1971): Hydrolysis of proteins with p-toluene-sulfonic acid:

Determination of tryptophan. 246:2842-2848. Martensen TJ (1982): Phosphatyrosine in proteins. J Biol Chem 257:9648-9652.

Matsubara H, Sasakr RM (1969): High recovery of tryptophan from acid hydrolysates of proteins. Biochem Biophys Res Connnun 35:175-181.

Mc Garvey B.D., Derivatization reactions applicable to pesticide determination by high- performance liquid chromatography. J. Chromatogr., B, 659, 243-257 (1994). Moore S, Stein WH (1963): Chromatographic determination of amino acids by the use of automatic recording equipment. Colowich SP, Kaplan NO (eds): "Methods in Enzymology."

New York: Academic Press pp 819-831.

Niall HD (1973): Automated Edman degradation. The protein sequenator. In Hirs CHW,

Timashoff SN (eds): "Methods in Enzymology." New York: Academic Press, Vol 27, pp 942-1010.

Niall H (1977): Automated methods for sequence analysis. In Neurath H, Hill RL (eds):

"The Proteins." 3rd edition, vol HI. New York: Academic Press, pp 179-238. Nice E C. Micropreparative high performance liquid chromatography of proteins and peptides: Principles and applications. Biopolymers 40(3).

1996. 319-341.

Penke B, et al. (1974): A new acid hydrolysis method for determining tryptophan in peptides and proteins. Anal Biochem 60:45-60.

Previero A, et al (1973): Solid phase sequential analysis: Specific linkage of acidic peptides by their car. boxyl ends to insoluble resins. FEBS Lp-tt 33:135138.

Previero A, Cobetti PMA (1977): "Solid Phase Methods in Protein Sequence Analysis. "

Amsterdam, New York, Oxford: Elsevier/North Holland Biomedical Press. Schaby AV, et al, (1976): Isolation and structure of somatostatin from porcine hypothalami.

Biochemistry 15:509-514.

Scott R.P.W., (1992) Chem. Soc. Reviews, Modern Liquid Chromatography, p. 137

Sher RT, et al. (1974): The amino acid sequence of porcine parathyroid hormone.

Biochemistry 13:1994-1999. Sherma J, (1994) Planar Chromatography, Anal. Chem., 66, 67R

Spackman DH et al. (1958): An automatic recording apparatus for use in the chromatography of amino acids. Anal Chem 30:1190-1202.

Swadesh JK, et al. (1984): Sodium sulfite as an antioxidant in the acid hydrolysis of bovine pancreatic ribonuclease A. Anal Biochem 141:397-401. Tarr GE (1977): Improved manual sequencing methods. Methods Enzymol 47:335-357.

Tsugita A, Scheffler JJ (1982): A rapid method for acid hydrolysis of protein with a mixture of trifluoroacetic acid and hydrochloric acid. Eur J Biochem 124:585588.

Waterfield M, Bridgen J (1975): In Perham RN (ed): "Instrumentation in Amino Acid

Sequence Analysis." London, New York, San Francisco: Academic Press, pps 41-71. Westall F, Hesser H (1974): Fifteen-minute acid hydrolysis of peptides. Anal Biochem

61:610- 613.

Wilson K (1986): "Precolumn Amino Acid Analysis Using Gas-Phase Hydrolysis and PITC

Derivalization. " Applied Biosystems, Inc. User's Bulletin.

Yeung E.S., Edit., Detectors for Liquid Chromatography, J. Wiley, NY, 362p, (1993) Yokote J, et al. (1986): Recovery of tryptophan from 25-minute acid hydrolysates of protein.

Anal. Biochem 152:245-249.

Claims

WHAT IS CLAIMED IS:

1. A method for determining the amino acid sequence of a polypeptide comprising: reacting the polypeptide with an Edman-type reagent that cleaves a terminal amino acid and binds to the amino acid to produce an Edman derivative ofthe amino acid, separating the products to resolve a peak consisting essentially ofthe Edman derivative, providing that the peak is radioactive by either (a) using an Edman-type reagent that is radiolabelled with a multiphoton emitting isotope prior to the derivatizing step, or (b) labeling the Edman derivative with a multiphoton emitting isotope after the derivatizing step, and fixing the resolution ofthe peak for a time sufficient to collect sufficient counts to permit identifying the amino acid from which the Edman derivative was derived using multiphoton detection techniques, identifying the amino acid from which the Edman derivative was derived using multiphoton detection techniques.

2. A method according to claim 1, further comprising quantifying the amino acid from which the Edman derivative was derived.

3. A method according to claim 1 , further comprising providing that the Edman-type reagent has a functional group suitable for radiolabelling with a multiphoton emitting isotope, protecting the functional group during derivatization, deprotecting the functional group after derivatization, and radiolabelling with a multiphoton emitting isotope before identifying the amino acid

4. A method according to claim 1 , wherein the Edman-type reagent is PITC or a homologue, and the Edman derivative is a phenylhydantoin-amino acid derivative.

5. A method according to claim 1, wherein the separation is chromatographic or electrophoretic

6. A method according to claim 1, wherein the separation is by HPLC and comprising freezing the column, eluting the column onto a solid absorbent medium, or tracking the peak with a multiphoton detection scanner having an aperture to increase the time the peak remains before the aperture.

7. A method according to claim 1, comprising scanning the peak repeatedly with increasing resolution.

8. A method for determining the amino acid composition of a polypeptide comprising: hydrolyzing the polypeptide, derivatizing the polypeptide with an Edman-type reagent that cleaves amino acids and binds to the amino acids to produce Edman derivatives ofthe amino acids, separating the products to resolve a plurality of peaks ofthe Edman derivatives, providing that the peak is radioactive by either (a) using an Edman-type reagent that is radiolabelled with a multiphoton emitting isotope prior to the derivatizing step, or (b) labeling the Edman derivative with a multiphoton emitting isotope after the derivatizing step, identifying the amino acids from which the Edman derivative was derived using multiphoton detection techniques, and quantifying the amino acid from which the Edman derivative was derived.

9. A device for sequencing proteins comprising: an HPLC column a movable spatially resolving multiphoton detector having an aperture oriented adjacent the HPLC column, a controller for the detector able to steer the detector to track movement of a peak within the column.

10. A reagent for use in determining the amino acid composition of a polypeptide comprising: a volatile aromatic compound of greater than about 90% purity capable of coupling to all types of amino acids, and of binding and cleaving to each amino acid in the polypeptide with a yield in excess of about 90%, and comprising either a stably-bound multiphoton emitting isotope or a protected functional group able to stably bind a multiphoton emitting isotope after deprotection, the degree of labeling being sufficiently high to permit detection ofthe amino acids of a few attomoles of polypeptide.

11. A method according to claim 10 , wherein the isotope is an iodine moiety bound at the 2, 3, 5, or 6 position, or two iodine moieties bound at the 2,3, 2,4, 2,5, 2,6, 3,4, 3,5, or

3,6 positions.

12. A method according to claim 10, wherein the reagent is an aromatic isothiocyanate.

13. A process for purifying the reagent of claim 10 comprising chromatographic separation ofthe reagent under control of multiphoton detection techniques.