Use of a proteome strategy for tagging proteins present at the plasma membrane


  • 1Reference plasma membrane 2D maps, together with protein sequences associated to the spots characterized in this work are available at the web site

*For correspondence (fax +33 499612 707; e-mail


A plasma membrane (PM) fraction was purified fromArabidopsis thalianausing a standard procedure and analyzed by two-dimensional (2D) gel electrophoresis. The proteins were classified according to their relative abundance in PM or cell membrane supernatant fractions. Eighty-two of the 700 spots detected on the PM 2D gels were microsequenced. More than half showed sequence similarity to proteins of known function. Of these, all the spots in the PM-specific and PM-enriched fractions, together with half of the spots with similar abundance in PM fraction and supernatant, have previously been found at the PM, supporting the validity of this approach. Extrapolation from this analysis indicates that (i) approximately 550 polypeptides found at the PM could be resolved on 2D gels; (ii) that numerous proteins with multiple locations are found at the PM; and (iii) that approximately 80% of PM-specific spots correspond to proteins with unknown function. Among the latter, half are represented by ESTs or cDNAs in databases. In this way, several unknown gene products were potentially localized to the PM. These data are discussed with respect to the efficiency of organelle proteome approaches to link systematically genomic data to genome expression. It is concluded that generalized proteomes can constitute a powerful resource, with future completion ofArabidopsisgenome sequencing, for genome-wide exploration of plant function.


The plant plasma membrane (PM) regulates the exchange of information and substances between the cell and its environment. Many of the functions are plant-specific, such as cell wall assembly and the response to pathogens, and involve PM-specific proteins. To date, however, only a few PM proteins have been characterized. Even in the model plant Arabidopsis thaliana, under 50 gene products have been identified among an estimated population of at least 500 polypeptides ( Masson & Rossignol 1995). In addition, these proteins fall into less than 20 multigene families of integral membrane proteins, mostly isoforms of a few extensively characterized transport systems ( Logan et al. 1997a ). Among other PM-resident proteins, there is almost no data available concerning peripheral proteins. Finally, nearly nothing is known about proteins with multiple locations in plants ( Robinson 1996), including those that are reversibly associated with the membrane after post-translational modifications ( Park et al. 1997 ;Xing et al. 1997 ).

Despite the availability of approximately 36 000 Arabidopsis ESTs in the databases, it is difficult to identify potential PM proteins. There is a clear need to develop methods to identify novel PM proteins and to link genes systematically to their products. In the last few years, approaches aimed at constructing links between proteins and the genome have become popular and led to the term ‘proteome’ to characterize the protein complement of a genome ( Wasinger et al. 1995 ). In yeast, with the availability of the complete set of genes, such proteome-based approaches immediately emerged as the most efficient way for systematic identification of gene products ( Shevchenko et al. 1996 ). Since the Arabidopsis genome sequencing will soon be complete, similar efficiency can be expected provided that the relevant proteome information has been made available.

In plants, relatively little effort has been devoted to proteomes and concerned total proteins to date ( Kamo et al. 1995 ). However, when performed on subcellular fractions, proteomes might allow identification of the intracellular location of unknown proteins. They may also reveal unexpected localizations of proteins already characterized elsewhere in the cell, as recently suggested in animals ( Fialka et al. 1997 ;Scianimanico et al. 1997 ). The aim of this work was to assess the feasibility of the proteome approach to tag PM proteins. The generated information was stored in an accessible database to constitute a resource allowing (i) direct comparison of homemade protein 2D gel maps with annotated reference gels1; (ii) recovery of the corresponding gene sequences; and (iii) identification of probable intracellular location.


A prerequisite for the construction of a PM proteome is the selection of procedures allowing for easy comparison between any PM-2D gels obtained in different laboratories. We selected a classical phase partitioning procedure to prepare membrane fractions enriched in PM from Arabidopsis leaves, and no specific washing procedure was used in order to address a population of PM proteins that is as comprehensive as possible. To improve and standardize the IEF step, commercially available immobilized pH gradients were selected. Finally, the ecotype Columbia, most frequently used to generate ESTs, was selected as biological material.

Approximately 700 spots were detected on a typical PM gel ( Fig. 1a). A comparison was performed between 2D gels of PM fractions and of the membrane supernatant, mostly containing cytosolic proteins and proteins released from broken organelles, in order to take into account a possible adsorption, or entrapment, of soluble proteins ( Fig. 1). Spots in the PM fraction were distributed into four classes on the basis of statistical analysis of their abundance in the different gels ( Table 1). One half were uniquely detected in the PM fraction. The other half were present in both the PM fraction and the supernatant. These were further classified according to their relative abundance in the two fractions as enriched, of similar abundance, or less abundant in the PM fraction. In order to test the relevance of this gel analysis-based classification, 82 proteins were sequenced ( Table 2;Fig. 2).

Figure 1.

Comparative analysis of 2D gels of PM and soluble proteins.

(a) PM proteins; (b) supernatant proteins.

Table 1.  Characteristics of the polypeptide population. Polypeptides simultaneously detected in the PM fraction and supernatant were distributed into classes according to statistical analysis of their abundance in the two. After protein sequencing, the percentage of polypeptides possibly located at the PM was extrapolated from the proportion of identified proteins for which PM location was previously reported
Classes of polypeptides
  Specific to
in PM
Similar abundance
in PM and supernatant
in supernatant
Gel analysis
 polypeptide number3555127032
Protein sequencing
 polypeptide sequenced1417447
 no close matches29%0%18%14%
 Possible PM location100%100%50%0%
Table 2.  Summary of proteins analyzed by 2D gel electrophoresis and protein sequencing
  1. Sequences for identified proteins are available from the database site ( For sequences of proteins with no close matches in databases the single letter code was used; the letter ‘X’ indicates the presence of an unidentified amino acid, ‘/’ refers to an indetermination between the two adjacent amino acids (the first one being the more likely), and ‘?’ indicates the uncertain identification of the preceding amino acid.

   PM-specific spots
721.65.7EST, A. thaliana[N38250]
26> 944.9EST, A. thaliana[T44457]
42> 944.6COR78, A. thaliana[L22567]
24123.05.5Glycine-rich protein, A. thaliana[S47408]
24621.65.8EST, A. thaliana[T76806]
28639.04.8ERD14, A. thaliana[D17715]
29753.94.7COR47, A. thaliana[X90959]
31131.05.0ERD14, A. thaliana[D17715]
57738.45.7Fructose diphosphate aldolase, A. thaliana[X53058]
77214.34.7Thioredoxin, A. thaliana[Z35476]
Spots enriched in PM
826.85.8Glutathione S transferase, A. thaliana[2262152]
1230.85.7V-ATPase subunit E, A. thaliana[X92117]
1331.35.3Carbonic anhydrase, A. thaliana[L18901]
1431.35.1Carbonic anhydrase, A. thaliana[L18901]
3240.25.6Protein disulfide isomerase, A. thaliana[2529680]
3627.65.5Glutathione S transferase ERD13, A. thaliana[D17673]
13042.15.7EST, A. thaliana[T43524]
15638.76.0Glyceraldehyde-3-phosphate dehydrogenase, A. thaliana[M64116]
19135.54.9Endomembrane-associated protein, A. thaliana[Y08061]
23527.35.7Glutathione S transferase, A. thaliana[2262152]
25617.85.6Actin depolymerizing factor, A. thaliana[U48939]
27429.05.1Carbonic anhydrase, A. thaliana[L18901]
28539.15.0Annexin, A. thaliana[X99224]
29330.25.1Carbonic anhydrase, A. thaliana[L18901]
29956.74.5Calreticulin, A. thaliana[U66343]
54634.35.1Annexin, A. thaliana[X99224]
76814.15.4EST, A. thaliana[T41851]
Spots of similar abundance in PM and supernatant
415.75.9Nucleoside diphosphate kinase, A. thaliana[X69373]
617.25.8Major latex protein, A. thaliana[X91961]
924.55.8Glutathione S transferase, A. thaliana[Z37698]
1026.95.9Glutathione S transferase, A. thaliana[2262152]
1126.66.1Glutathione S transferase ERD13, A. thaliana[D17673]
1942.49.0Isocitrate dehydrogenase, Nicotiana tabacum[X77944]
2038.86.0Glyceraldehyde-3-phosphate dehydrogenase, A. thaliana[M64116]
2476.45.9Methionine synthase, Solenostemon scutellarioides[Z49150]
2952.65.5Enolase, A. thaliana[119350]
3340.55.5Protein disulfide-isomerase, A. thaliana[2529680]
3428.95.1Triose phosphate isomerase, A. thaliana[U02949]
4575.04.6Methionine synthase, Catharanthus roseus[X83499]
6862.35.4Phosphoglycerate mutase, A. thaliana[2160168]
7062.75.2Phosphoglycerate mutase, A. thaliana[2160168]
7869.44.9Heat shock protein 70, A. thaliana[Z18432]
13147.85.6Calreticulin, A. thaliana[U66345]
14042.75.5S-adenosylmethionine synthetase, A. thaliana[M55077]
14342.35.1Actin, A. thaliana[U27811]
15737.86.1Glyceraldehyde-3-phosphate dehydrogenase, A. thaliana[M64116]
16337.55.7Malate dehydrogenase, A. thaliana[2341034]
26616.34.7EST, A. thaliana[N37338]
27214.64.8Profilin, A. thaliana[U43323]
27622.04.6EST, A. thaliana[Z29954]
27919.14.7EST, A. thaliana[N65016]
28019.74.8EST, A. thaliana[Z29954]
28437.05.0EST, A. thaliana[H77174]
28756.04.5Calreticulin, A. thaliana[U27968]
29034.74.8EST, A. thaliana[T22397]
29560.84.8Protein disulphide isomerase?, Ricinus communis[U41385]
29856.94.7COR47, A. thaliana[X90959]
30717.34.6EST, A. thaliana[T21401]
31333.24.7EST, A. thaliana[T04262]
32728.44.5EST, A. thaliana[Z33718]
37327.85.6Ascorbate peroxidase, A. thaliana[X59600]
54735.35.1Acidic ribosomal protein, A. thaliana[2088654]
54929.55.7EST, A. thaliana[Z33676]
Spots enriched in supernatant
514.75.8Ribulose bisphosphate carboxylase small subunit, A. thaliana[X14564]
1637.05.8Malate dehydrogenase, Zea mays[AF007581]
3526.15.2EST, A. thaliana[T42745]
9649.95.6Ribulose bisphosphate carboxylase large subunit, Allium cepa[D38294]
22828.85.3Triose phosphate isomerase, A. thaliana[U02949]
26915.35.7Ribulose bisphosphate carboxylase small subunit, A. thaliana[X14564]
Figure 2.

Arabidopsis leaf PM proteome.

Sequenced spots ( Table 2) are identified by their spot number; circles, squares, crosses, ovals: spots classified as, respectively, PM-specific, enriched in PM, equally distributed in PM and supernatant, and enriched in supernatant.

For the first class, i.e. putative PM-specific proteins, a large proportion of novel or functionally uncharacterized proteins was found. These either (i) matched with products of genes of unknown function (dehydrins, COR and ERD genes;Horvath et al. 1993 ;Kiyosue et al. 1994 ) some of which were known to be located at the PM ( Danyluk et al. 1998 ); or (ii) were related to unidentified ESTs; or (iii) gave no close matches in the databases. Only three spots matched to well-characterized proteins: a glycine-rich protein, a fructose diphosphate aldolase, and a thioredoxin. The glycine-rich protein is a member of a family of proteins characterized in various plant species and thought to be located at the cell wall/membrane interface ( Condit 1993;de Oliveira et al. 1990 ). The aldolase is a glycolytic enzyme; however, in addition to a cytosolic location, a phosphorylation-promoted association of the enzyme with the PM has been described in animals ( Hardin & Paul 1992;Kaklij & Kelkar 1983). The thioredoxin belongs to a large family of proteins exhibiting various subcellular locations. Although most of them are chloroplastic or cytosolic, one was located at the PM in soybean ( Shi & Bhattacharyya 1996). These results are therefore consistent with proteins in this category having a PM location.

Of 17 sequenced proteins from the second class (i.e. enriched in the PM fraction), 14 showed significant homology with proteins present in databases, the remaining ones matching with sequences translated from ESTs. Three proteins were identified as different glutathione-S-transferases (GSTs). GSTs constitute a large family of proteins, some of them are known to be located at the PM in Arabidopsis and in animals ( Horbach et al. 1993 ;Zettl et al. 1994 ). Four proteins matched with carbonic anhydrases (CAs). In plants, CAs have been mainly described in the cytosol and in chloroplasts, but they are presumed also to be located at the PM, as is well-established in algae and in animals ( Fisher et al. 1996 ;Sültemeyer et al. 1993 ). Two proteins were characterized as annexins. Depending on calcium concentration, these proteins are able to interact with PM phospholipids and are also suspected to act as ion channels ( Moss 1995). In plants, they were localized at the PM in cotton fiber and Bryonia dioica ( Andrawis et al. 1993 ;Thonat et al. 1997 ). One spot matched with a V-ATPase subunit. This type of ATPase is usually thought to be located at the tonoplast in plants. However, it has also been immunolocalized to the pea PM, in agreement with the location observed in many eukaryotic cells ( Finbow & Harrison 1997;Robinson et al. 1996 ). One spot was identified as a protein disulfide-isomerase (PDI). Though largely localized in the ER, PDIs were also demonstrated to be present at the PM in various animal cell types where they are involved in the processing of receptors ( Couet et al. 1996 ;Terada et al. 1995 ). Spot 156 showed sequences identical to a glyceraldehyde-3-phosphate dehydrogenase. As for the aldolase above, this glycolytic enzyme was demonstrated to be reversibly associated with PM in some animal cells through the action of proteases ( Hardin & Paul 1992;Hsu & Molday 1990). Spot 191 matched an Arabidopsis protein described as an endomembrane associated protein. However, a homologous protein was identified at the tobacco PM ( Logan et al. 1997b ). One spot was identified as an actin depolymerizing factor. In animals, this protein was localized both in the cytosol and at the PM in various cell types, the reversible association with the membrane being controlled by dephosphorylation ( Abe et al. 1996 ;Suzuki et al. 1995 ). The last spot identified in this class matched with a calreticulin. Although mainly located at the ER various cellular locations are known for this protein in animals including at the PM, and a calreticulin-like protein was recently reported in a tobacco PM fraction ( Droillard et al. 1997 ;Gray et al. 1995 ). Therefore, all the characterized proteins in this category have previously been described at the PM in various organisms.

For the third class, corresponding to proteins with similar distribution between the PM fraction and supernatant, 44 spots were sequenced. Eighteen spots showed no close matches in databases or were related to ESTs. Among identified spots, nine corresponded to proteins also present in the second class and for which some evidence of a possible PM location is available: three GSTs, two glyceraldehyde-3-phosphate dehydrogenases, two calreticulins, and two PDIs. Spot 4 matched with nucleoside diphosphate kinases. This enzyme is involved in various transduction processes at the PM in animals and was characterized in pea PM ( Bominaar et al. 1993 ;Hamada et al. 1996 ). Two spots were identified, respectively, as actin and profilin. Although the nature of the interactions between microfilaments and the PM is not clear in plants, some evidence argues for a connection between actin filaments and the PM ( Lloyd et al. 1996 ). The phospholipid- and actin-binding protein profilin has been demonstrated to be reversibly translocated to the PM of eukaryotic cells according to the abundance of phosphatidylinositol 4,5-bisphosphate ( Ostrander et al. 1995 ). Spot 78 matched with a heat shock protein hsp70. This chaperone was hypothesised to be involved in protein folding/assembly on the PM and was recently shown to be expressed at the surface of specialized animal cells ( Mamelak & Lingwood 1997;Shi et al. 1995 ). For the 13 other proteins, no evidence favouring a putative PM localization is available, with the possible exception of spot 373. This spot was identified as a member of the ascorbate peroxidase family in which the precise location of two membrane-bound isotypes remains unclear ( Jespersen et al. 1997 ). Half the characterized proteins in this category have therefore previously been described at the PM.

In contrast, all identified spots from the last class (i.e. proteins more abundant in the supernatant) are soluble proteins with no evidence of a functional interaction with the PM.


In recent years, general proteomes have been constructed for numerous organisms with high success in protein identification. In this work, we have begun to establish the proteome of Arabidopsis PM. Of 700 spots found in a PM-enriched fraction, sequence of 82 was obtained; 29 of these proteins have not been previously identified. In generating an organelle proteome, it is necessary, however, to validate the proposed location of the proteins.

Protein characterisation according to gel analysis and sequence data

Owing to the low molecular knowledge of the PM, systematic approaches are likely to identify mainly previously unknown proteins or proteins already characterized at another location. We assessed the specificity by comparing the protein composition of the PM preparation with a microsomal supernatant, and classified the proteins according to their unique presence or relative abundance in the fractions ( Table 1). Many of the proteins that could be identified by sequencing have already been characterized at the PM in various organisms. Moreover, the proportion of these putative PM proteins was related to their enrichment in the Arabidopsis PM fraction: there is previous evidence for a PM localization for all the identified PM-specific and PM-enriched proteins, but there is no such evidence for half the proteins of similar abundance in both fractions and for all the proteins enriched in the supernatant fraction. Therefore, the gel-based characterization of PM proteins is supported by protein identification.

Validity of protein identification

The usual tools were used to query the databases ( Altschul et al. 1990 ). For identified spots, an average match of 95% identity, over an average sequence length of 14 amino acids, was found. The probability that these scores were obtained by chance is low, but two points deserve further discussion. For one-third of the identified spots, not all of the microsequences obtained were found to match the database entry. This is particularly likely for spots matching with partly sequenced ESTs. The incomplete matches may also mean that these spots correspond to isoforms not present in the databases. However, it cannot be ruled out that different proteins co-migrated in the same spot, although in no case did the sequences obtained in this work identify two proteins in one spot. Another striking feature is that different spots matched with the same protein type. This might result from protein degradation. However, the relative molecular weights measured on gels is lower than that calculated for the corresponding translated protein (by less than 4%) for only two spots (spots 131 and 546). This does not preclude any degradation but suggests that, if occurring, it would concern a minor proportion of identified proteins. Therefore, the presence of multiple spots with similar sequences probably reflects the presence of isoforms and/or of post-translational modifications.

The PM, a membrane rich in peripheral proteins

Most proteins identified here as putative PM-proteins are peripheral proteins. Although this might be an unexpected result, similar conclusions have been drawn from the only systematic approach of the plant PM published to date ( Shi et al. 1995 ). Immunoscreening of an expression library allowed these authors to identify mainly peripheral and multi-location proteins, including three proteins identified again in the present work. Together, these data suggest that such proteins account for a substantial part of the total population and can play a previously poorly recognized role in PM function. On the other hand, due to the high abundance of peripheral proteins, integral proteins are likely to constitute minor proteins that were not analyzed here due to limitations in the amount available. The under-representation of very hydrophobic proteins may be due also to the presently available proteome technology itself. Alternative detergents have been recently synthesised and shown to improve the recovery of some integral proteins on 2D gels ( Chevalet et al. 1998 ). Once commercially available and proven at a large scale, such detergents may be helpful in establishing complementary standardized procedures allowing the construction of more comprehensive proteomic resources.

The PM, a source of new and plant specific proteins

By extrapolation of the sequencing results ( Table 1), the population of putative PM proteins can be estimated to amount to approximately 550 polypeptides. Hence, the membrane fractions usually taken as PM-enriched could contain a substantial proportion (about 20%) of non-PM proteins. However, no definite conclusion can yet be drawn concerning the actual size of the PM protein population since the extrapolation made may constitute an overestimation, whereas some proteins are probably poorly represented. A more meaningful extrapolation concerns the type of proteins in this population. Assuming that the proportions of identified proteins are representative, the population of PM proteins analysable by a proteome approach would correspond to approximately 200 proteins already partially functionally characterised, and this is approximately four times the number of PM proteins identified to date by dedicated approaches. In addition, under the same assumptions, approximately 220 of the proteins not identified at the function level would presently identify ESTs. Finally, the remaining approximately 130 proteins would correspond to orphan- and possibly plant-specific proteins, in as much as they do not yet match any database entries.

Few plant proteomes have so far been initiated. Together with a previous report ( Kamo et al. 1995 ), our results demonstrate that they constitute a powerful way to (i) characterize new proteins (13 proteins with no close matches in databases in this work); (ii) identify the products of unidentified genes (16 proteins matching with ESTs); or (iii) propose a location for specific classes of genes (7 ERD-and COR-related proteins). Therefore, nearly half of this sequencing effort resulted in the tagging of previously uncharacterized proteins. In the case of organelle proteomes, a new dimension is introduced a priori by the membrane dynamics, resulting in various locations for the same protein type. Our data constitute the first report concerning a plant organelle proteome. They show that the expected increase in complexity is effectively seen at the proteome level, and suggest that comparative gel analysis can offer a convenient guide for subsequent protein analysis as recently proposed in animals ( Fialka et al. 1997 ;Scianimanico et al. 1997 ). A high output can therefore be expected from organelle proteomes, and a general proteome strategy connecting proteomes of individual organelles can now be proposed for genome-wide exploration of plant function.

Experimental procedures

Plant culture, plasma membrane purification and protein solubilization

Plants were cultivated under controlled conditions (light intensity of 150 μE m–2 s–1, 8 h day, 70% relative humidity, 20°C constant, temperature). Vegetative plants were collected 50 days after sowing.

A microsomal fraction was obtained from a leaf homogenate ( Santoni et al. 1990 ) and the resulting supernatant was used for the comparison between membrane and soluble proteins. A PM-enriched fraction was purified from microsomes by phase partitioning ( Larsson et al. 1990 ). The sensitivity of the Mg-ATPase activity to vanadate, oligomycin and KNO3 was used as markers of, respectively, PM, mitochondria and tonoplast, and the presence of Golgi membranes was assessed by the IDPase activity ( Santoni et al. 1990 ). The activities sensitive to vanadate, oligomycin and KNO3 amounted, respectively, to 83%, less than 1% and 13% of the total ATPase activity, whereas the IDPase activity represented 4% of the total ATPase activity (n = 15). These features are similar to those obtained for various other PM-enriched fractions prepared by phase partitioning ( Larsson et al. 1990 ).

Proteins were solubilized with 4% SDS and precipitated with cold acetone. The dried pellet was resuspended in 9 m urea, 0.5% Triton X-100, 20 m m DTT, 1.2% pharmalytes (pH 3–10) and 4% CHAPS. Protein amounts were estimated using a modified Bradford procedure ( Stoscheck 1990).

2D gel electrophoresis

For analytical purposes, IEF was performed on immobilized pH gradients (IPG; non-linear pH 3–10 gradient, Pharmacia). Proteins (50 μg) were loaded at the cathodic side (Multiphor, Pharmacia). The following running conditions were used: from 0 to 300 V in 1 min, 300 V for 30 min, from 300 V to 3500 V in 3 h, 3500 V for 3 h, 5000 V until a total of 85 kVh was reached. IPG gel strips were then incubated at room temperature for 10 min in 6 m urea, 30% (w/v) glycerol, 2% SDS, 2% DTT, 0.05 m Tris–HCl pH 6.8 and a trace of bromophenol blue. Another equilibration step was performed for 5 min in the same solution, with 2.5% iodoacetamide instead of DTT. The second dimension (PAGE-SDS) was carried out on homogenous 12% T gels (Protean II, BioRad). After electrophoresis at 20 mA for 1 h, and then at 40 mA, gels were silver-stained ( Santoni et al. 1997 ).

For preparative purposes, homemade linear IPG (pH 4–8) were used. Proteins (1.8 ml per gel, approximately 4 mg ml–1) were loaded onto the surface of dried gels, and gels were allowed to rehydrate with the sample overnight at 25°C. Electrophoresis was as above until a total of 130 kVh was reached. For PAGE-SDS, an additional 2 cm stacking gel was used. Gels were stained with Coomassie blue R-250.

Analysis of two-dimensional gels

The 2D gels were scanned with a laser scanner (ImageMaster Desk Top scanner) and analyzed using BioImage software according to Santoni et al. (1997) . Three sets of independent experiments were performed to detect qualitative and quantitative differences between PM and supernatant. Spots were taken as showing qualitative variations when they were systematically present (respectively absent) in the three analyzed PM 2D gels and absent (respectively present) in the three supernatant 2D gels. These spots were referred as specific, respectively, the PM and the supernatant. To analyze quantitative differences between spots present both in PM and supernatant, the abundance of each spot in PM gels was normalized to the total absorbance of spots matching with a reference PM gel, and intermediate gels of an equal mixture of PM and supernatant proteins were used to allow normalization to the reference PM gel. The normalized amounts of protein were compared using variance analysis ( SAS 1988). Accordingly, these spots were classified as enriched in one of the two, or as present with similar abundance in PM and supernatant.

Protein sequencing

Coomassie blue-stained spots were digested by trypsin, and peptides were purified by RP-HPLC ( Matsudaira 1993). After transfer of peptides to protein sequencer (LF 3000, Beckman; ABI models 477 A and 490), PTH-amino acids were identified with on-line HPLC.


This work was funded by the European Community BIOTECH programme (contract BIO4-CT95–0147).