Fiber type-specific expression of major proteolytic systems in fast- to slow-transforming rabbit muscle

Karim R. Sultan, Bernd T. Dittrich, Elmi Leisner, Nina Paul, Dirk Pette


The present study investigates the role of two major proteolytic systems in transforming rabbit and rat muscles. The fast-to-slow transformation of rabbit muscle by chronic low-frequency stimulation (CLFS) induces fast-to-slow transitions of intact, mature fibers and replacement of degenerating fibers by newly formed slow fibers. Ubiquitination, an indicator of the ATP-dependent proteasome system, and calpain activity were measured in homogenates of control and stimulated extensor digitorum longus muscles. Calpain activity increased similarly (∼2-fold) in stimulated rat and rabbit muscles. CLFS had no effect on protein ubiquitination in rat muscle but led to elevations in ubiquitin protein conjugates in rabbit muscle. Immunohistochemistry was used to study the distribution of μ-calpain and m-calpain and of ubiquitinated proteins in myosin heavy chain-based fiber types. The findings suggest that both proteolytic systems are involved in fiber transformation and replacement. Transforming mature fibers displayed increases in μ-calpain and accumulation of ubiquitin protein conjugates. The majority of these fibers were identified as type IIA. Enhanced ubiquitination was also observed in degenerating and necrotic fibers. Such fibers additionally displayed elevated m-calpain levels. Conversely, p94, the skeletal muscle-specific calpain, decayed rapidly after stimulation onset and was hardly detectable after 4 days of CLFS.

  • calpains
  • chronic low-frequency stimulation
  • fiber type
  • ubiquitination

chronic low-frequency stimulation (CLFS) is an established model for inducing fast-to-slow transformation of skeletal muscle (19-21). This transformation process encompasses a remodeling of the myofibrillar apparatus by sequential exchanges of fast with slower myofibrillar protein isoforms. A prerequisite of the insertion of newly synthesized protein isoforms into the sarcomere is the excision of isoforms no longer synthesized. Therefore, proteolysis has been suggested to be a limiting step in fast-to-slow fiber type transitions (33). Indeed, fast-twitch rat muscle exposed to CLFS exhibits elevated calpain activity, and this increase was shown to be related to changes in the intracellular distribution of μ-calpain (32).

The CLFS-induced fast-to-slow conversion encompasses, in addition to fiber type transitions, a pronounced decrease in whole muscle cross-sectional area resulting from fiber atrophy (18) but not from fiber loss (17). Fiber type transitions due to enhanced neuromuscular activity potentially involve two different processes: 1) transformation of uninjured fast into slow fibers and 2) exchange of deteriorated fast fibers by newly formed slow fibers (13). All available experimental data indicate that fast-to-slow transitions in rat muscle are solely based on transformation of intact, mature fibers (4, 23). On the other hand, rabbit fast-twitch muscle exposed to CLFS displays pronounced signs of fiber degeneration and regeneration. As previously shown, ∼15–20% of the fast fiber population degenerates and is replaced by slow fibers originating from proliferating satellite cells (12-14).

We assumed that the proteolytic processes involved in fiber type transitions and in fiber type replacement differ with regard to the quality, extent, and location of the proteolytic systems involved. In this context, the major proteinase families of interest are the calpains (2) and the ATP-dependent ubiquitin-related proteasome system (11). An additional proteinase of interest is the muscle-specific calpain 3, p94 (9). The observation that fast-twitch glycolytic fibers preferentially undergo degeneration and regeneration in low-frequency stimulated rabbit muscle (12-14) points to fiber type-specific patterns of proteolysis. Because fiber necrosis is accompanied by an invasion of mononucleate nonmuscle cells, additional proteolytic systems involved in phagocytosis must be considered (10).

The present study of rabbit fast-twitch extensor digitorum longus (EDL) muscle exposed to CLFS investigates calpain activity, changes in the expression of p94, the ubiquitination of muscle proteins, and its correlation with specific fiber types during the early period of transformation. For this purpose, EDL muscles were continuously stimulated at low frequency for 1, 4, and 8 days and compared with the contralateral unstimulated muscles as well as the slow-twitch soleus (SOL) muscle. For comparison, studies were also performed on low-frequency stimulated rat muscles. Calpain activity was measured in muscle homogenates using a specific fluorogenic substrate, p94 content and protein ubiquitination were determined by immunoblotting, and fiber type distribution patterns were assessed by immunohistochemistry with monoclonal antibodies specific to myosin heavy chain (MHC) isoforms MHCI, MHCIIa, and MHCIId. Desmin and dystrophin distribution were studied to define muscle fiber integrity.


Animals, CLFS.

Muscles were obtained from adult male rabbits (White New Zealand strain) and adult male Wistar rats. Fast-twitch EDL and tibialis anterior (TA) muscles were chronically stimulated at low frequency (10 Hz, 0.2-ms impulse width, 24 h/day) via electrodes implanted laterally to the peroneal nerve of the left hindlimb for 1 day (rabbit: n = 2), 2 days (rat: n = 4), 4 days (rabbit: n = 4; rat: n = 4), and 8 days (rabbit: n = 2; rat: n = 4) as previously described (27). Animals were killed under anesthesia, and stimulated EDL and TA muscles (left hindlimb) and contralateral control muscles including SOL (right hindlimb) were excised, frozen in melting isopentane (−159°C), and stored in liquid nitrogen.

Preparation of muscle extract for calpain assay.

The frozen muscles were pulverized under liquid nitrogen in a steel mortar. The following procedures were performed at 4°C. Muscle powder was homogenized (polytron homogenizer; Kinematica, Luzern, Switzerland) at 10,000 rpm at intense cooling with an ice-salt mixture in a fivefold volume of 20 mM Tris · HCl buffer (pH 7.4) that contained 5 mM EDTA, 5 mM EGTA, and 1 mM dithiothreitol (DTT, Sigma) as well as 10 μg/ml Pefabloc [4-(2-aminoethyl)benzylsulfonylfluoride hydrochloride; Roth, Karlsruhe, Germany] and 10 μg/ml pepstatin A (Calbiochem) as proteinase inhibitors. Three homogenization steps, each lasting 60 s, with intervals of 3 min were used. Extracts were separated from debris by 10-min centrifugation at 1,000 g. Protein concentration was determined by the Bradford method (Bio-Rad protein assay) with BSA as a standard.

Calpain assay in the presence of endogenous calpastatin.

Calpain was measured using a modification of the assay of Edelstein and coworkers (5).N-succinyl-Leu-Tyr-7-amido-4-methylcoumarin (SLY-AMC) served as a substrate for calpain (25). A stock solution of 50 mM SLY-AMC was prepared in dimethylsulfoxide and stored at −20°C. The following procedure was used for measuring calpain activity in muscle extracts: 30 μl muscle extract was incubated for 10 min at 37°C in a buffer solution (pH 7.4) that contained 20 mM Tris · HCl, 5 mM CaCl2, 1 mM DTT, 10 μg/ml Pefabloc, and 10 μg/ml pepstatin A. After addition of 5 μl of the substrate solution, buffer was added to adjust the volume of the assay to 2 ml. Fluorescence of the liberated AMC was monitored in a Perkin Elmer fluorometer for 15 min at 37°C (excitation 380 nm, emission 460 nm). Activity was expressed as arbitrary units per minute of incubation time per milligram of muscle protein.

Immunoblot for ubiquitin protein conjugates and p94.

Muscle powder was homogenized in 10 vol of reducing sample buffer and heated for 10 min at 90°C. After separation on a 7.5% (for ubiquitination) or 10% (p94) SDS polyacrylamide gel, proteins were blotted on a nitrocellulose membrane (Schleicher and Schüll) according to the method of Towbin et al. (34). Equal amounts of proteins were confirmed by staining with Ponceau S and documented for densitometric standardization. The membrane was blocked for 2 h with 5% fat-free milk powder and 2% BSA in 20 mM Tris · HCl (pH 7.6), 137 mM NaCl, and 0.1% Tween 20. The primary mouse monoclonal antibody to ubiquitin and ubiquitinated proteins (SC-8017; Santa Cruz Biotechnology) was diluted 1:2,000 in blocking solution, and membranes were incubated for 2 h at room temperature. The primary monoclonal mouse antibody to calpain 3 (p94, clone Calp3d/2C4; Novocastra Laboratories) was diluted 1:20 in blocking solution without milk powder, and membranes were incubated overnight at 4°C. Peroxidase-coupled secondary antibodies were applied, and detection was performed with the enhanced chemiluminescence Western blot detection reagent (Amersham). After detection, blots were washed and reincubated with a monoclonal anti-desmin mouse antibody (Sigma D1033). The intensity of the bands was estimated by integrating densitometry (ScanPack software; Biometra, Göttingen, Germany). The degree of ubiquitination was estimated by evaluating ubiquitinated proteins of the whole lane. Both values were standardized with the Ponceau S intensity.


Nine-micrometer-thick frozen sections were air dried, washed twice in phosphate-buffered saline that contained 0.1% Tween 20 (PBS-t, pH 7.4), and incubated for 15 min in 3% H2O2 in methanol. Sections were blocked for 2 h in PBS-t that contained 2% BSA and 10% horse serum. For staining of MHC isoforms, the primary mouse monoclonal antibodies were diluted in blocking solution: anti-MHCI (7HCS15) 1:30; anti-MHCI and MHCIIa (7HCS11) 1:400; and the BF-35 antibody (that recognizes all MHC isoforms except MHCIId/x) 1:1,000. The specificities of the 7HCS15, 7HCS11, and BF-35 antibodies have been described (26). Monoclonal antibodies were used for staining μ-calpain (MA3–940), m-calpain (MA3–942), calpastatin (MA3–945) obtained from ABR (Alexis, Läufelingen, Switzerland), desmin (clone DE-B-5, Roche), and dystrophin (clone Dy4/6D3; Novocastra) as controls for muscle fiber integrity. After overnight incubation with the primary antibodies, sections were washed and reacted for 30 min with biotinylated secondary antibodies. Thereafter, sections were washed, incubated for 30 min with a biotin-avidin-horseradish peroxidase complex (Vectastain ABC reagent; Vector Laboratories, Burlingame, CA), and washed and reacted for 5–10 min with staining solution (3,3′-diaminobenzidine tetrahydrochloride peroxidase substrate kit SK-4100; Vector).

For all immunohistochemical stainings, contralateral control and stimulated muscles were mounted together to cut cross sections of identical thickness and to perform all reactions under identical conditions. The comparison of sections from control and stimulated muscles on the same slide facilitated the detection of specifically stained fibers in the stimulated muscle. This was especially important in the case of ubiquitination, which was also present, although at lower levels in the controls. Immunohistochemical stainings were evaluated by examining four to six randomly selected fields (∼150 fibers each) for each condition.

Immunoblot and calpain activity results are given as means ± SD. Student's t-test was used to determine whether differences existed between results from different muscles or experimental conditions. The acceptable level of significance was set at P < 0.05.


The major purpose of the present study was to elucidate the role of the two major cytosolic proteolytic systems in transforming rabbit muscle. In view of our previous findings on rat muscle, in which major effects were observed after 4 days of CLFS (32), the present study primarily focuses on this time point. Additional studies were performed on muscles exposed to shorter or longer time periods of CLFS.

Ubiquitin protein conjugates in transforming muscle.

Electrophoretically separated proteins from control and stimulated muscles displayed multiple signals as detected by a monoclonal antibody recognizing ubiquitinated protein conjugates (Fig.1). The intensity of the signals increased significantly (∼3-fold) in 4-day stimulated rabbit muscles. Conversely, CLFS was without effect on rat EDL muscles, because 4- and 8-day stimulated muscles exhibited (qualitatively and quantitatively) signals for ubiquitinated protein conjugates similar to their contralateral controls. The failure to detect changes in ubiquitination between controls and stimulated rat muscles was confirmed by densitometric evaluations of the blots (Fig. 1) and immunohistochemistry (data not shown).

Fig. 1.

Densitometric evaluation of immunoblot analysis for ubiquitin-protein conjugates in low-frequency stimulated extensor digitorum longus (EDL) muscles from rabbit (open bars) and rat (solid bars). Values (means ± SD, n = 3) are expressed as relative intensities referred to the unstimulated, contralateral muscles (n = 3). A typical immunoblot is shown (inset). C, contralateral control; S, stimulated; d, days; CLFS, chronic low-frequency stimulation. *P < 0.05.

Fiber type distribution of ubiquitin protein conjugates.

In view of the earliest significant increase in ubiquitin protein conjugates, immunohistochemistry was performed on 4-day stimulated muscles. Inspection of cross sections revealed that stimulated muscles reacted stronger compared with the basal staining level in the control muscles (Fig. 2). In general, the stimulated muscles displayed two morphological patterns: 1) a checkerboard-like staining of fibers morphologically judged as intact (Fig.2 a) and 2) more intensely stained fibers within areas abundant in mononucleated infiltrates (Fig. 2 b). Such areas were observed in all stimulated muscles. In addition, stimulated muscles contained enlarged, vacuolated fibers with rounded shape and reduced levels of ubiquitination (Fig. 3 B). Occasionally, these fibers also displayed infiltrations with mononucleated cells. Such fibers were not seen in the control muscles.

Fig. 2.

Immunohistochemical detection of ubiquitinated fibers in control and 4-day stimulated rabbit EDL muscles. In general, stimulated muscles display 2 staining patterns, a checkerboard-like staining of fibers morphologically judged as intact (a), and intensely stained fibers within areas that contain infiltrations by mononucleated cells (b). Bar, 100 μm.

Fig. 3.

Myosin heavy chain (MHC)-based immunohistochemical identification of ubiquitinated fibers in 4-day stimulated rabbit EDL. A: representative area of injured muscle fibers. Fibers 1 and2 display enhanced staining for ubiquitinated protein conjugates (Ub-PC) and are identified as type IIA fibers by their positive reaction with an antibody specific to MHCI and MHCIIa and their negative reaction with an antibody specific to MHCI. Fibers 3 and 4, staining moderately or negative for Ub-PC, respectively, are identified as type IID fibers because they do not react with an antibody recognizing all MHC isoforms except for MHCIId.B: representative area of mostly intact fibers. *Fibers that display enhanced ubiquitination are identified as type IIA because of their reaction with the 7HCS11 antibody specific to MHCIIa and MHCI (bottom) without being stained by the 7HCS15 antibody specific to MHCI (not shown). Two enlarged ubiquitin-negative fibers with rounded shape and vacuolization are marked by arrows. Bar, 50 μm.

The quantitative evaluation of enhanced ubiquitination in MHC-based fiber types revealed that ∼80% of the intensely stained fibers in 4-day stimulated muscles were type IIA and 15% were type IID (Fig. 3, Table 1). However, only 34% of all type IIA fibers displayed enhanced staining for ubiquitin protein conjugates (Table 1).

View this table:
Table 1.

Fiber type distribution of enhanced ubiquitination in 4-day stimulated rabbit EDL muscle

Calpain activity and fiber type distribution.

Total calpain activity was assessed in muscle homogenates, i.e., in the presence of its endogenous inhibitor calpastatin. Similar to previous results on low-frequency stimulated rat muscle, CLFS for 4 days induced an ∼2.5-fold increase in calpain activity in rabbit muscle (Fig.4). Calpain activity measurements on 8-day stimulated rabbits revealed similar increases (data not shown). It cannot be excluded that the increase of total calpain activity, as determined in muscle homogenates, resulted, at least to some extent, from calpain activity of nonmuscle cells, i.e., infiltrates observed in the stimulated muscles. This possibility was examined by immunohistochemical detection of calpains in muscle cross sections (Figs. 5 and6). Stainings for μ-calpain and m-calpain identified fibers expressing both calpains at higher levels compared with their levels in control muscles (Fig. 5). These fibers also displayed elevated calpastatin levels. As could be expected (30, 31), elevated m-calpain levels were mainly detected in injured and necrotic fibers. Conversely, elevated μ-calpain levels were mainly seen in type IIA fibers, but not in necrotic fibers (Fig.6). According to quantitative evaluation of MHC-based fiber types, >90% of the fibers with elevated μ-calpain were type IIA. However, not all type IIA fibers displayed elevated μ-calpain content. Their fraction encompassed ∼50% of the type IIA population.

Fig. 4.

Total calpain activity in homogenates of low-frequency stimulated EDL muscles from rabbit (open bars) and rat (solid bars). Calpain activity was assessed by the cleavage of the calpain-specific fluorogenic substrateN-succinyl-Leu-Tyr-7-amido-4-methylcoumarin (SLY-AMC). Values (means ± SD, n = 3) are expressed as relative activities referred to the unstimulated, contralateral muscles (n = 3). *P < 0.05, **P < 0.01.

Fig. 5.

Serial sections of contralateral control (left) and 4-day stimulated (right) rabbit EDL muscles immunohistochemically stained for μ-calpain (C-I), m-calpain (C-II), and calpastatin (CST). No fiber type-specific distribution can be detected for the 2 calpains and calpastatin in the control muscle. Bar, 100 μm.

Fig. 6.

Immunohistochemical demonstration of μ-calpain, m-calpain, and calpastatin in serial sections of 4-day stimulated rabbit EDL muscle fibers identified according to their MHC complement. Fiber integrity was checked by immunohistochemistry for dystrophin and desmin and morphologically judged by hematoxylin-eosin (HE) staining. Note that calpastatin staining correlates with enhanced expression of μ- and m-calpain in the same fibers. Fibers positive for μ-calpain are identified according to their staining with the various MHC-specific antibodies (see also legend to Fig. 3) as type IIA. Necrotic fibers (arrows) display high levels of m-calpain but do not react for μ-calpain. Bar, 50 μm.

Skeletal muscle-specific calpain p94.

To investigate p94 expression levels by immunoblotting in stimulated and control muscles, protein amounts applied to the gels were adjusted to equal desmin contents. Desmin was chosen as a reference because of its rapid loss from injured fibers (6). This excluded the possibility that changes in p94 in the stimulated muscles resulted from fiber necrosis (Fig. 7). A pronounced decay of p94 was observed. The reduction of the p94 content, therefore, most likely occurred in uninjured fibers. The 4-day stimulated rabbit muscles contained only 10% of the p94 level in control muscles (Fig.8). Similar reductions in p94 level were observed in 1-day stimulated muscle, which made it unlikely that this rapid decline related to the fast-to-slow conversion. Moreover, control fast-twitch TA and slow-twitch SOL muscles displayed similar p94 levels (Fig. 8, inset). This result differs from a recent report on p94 content that was lower in slow than in fast glycolytic porcine fibers (8). Obviously, species-specific differences exist with regard to the distribution of p94 in fast and slow fibers.

Fig. 7.

Immunoblot analysis of skeletal muscle-specific calpain p94 (∼95 kDa) in contralateral control (C) and stimulated (Stim) EDL muscles from rabbit. After detection of p94, the blot was washed and reacted with a monoclonal antibody for desmin as a quantitative control for fiber integrity.

Fig. 8.

Densitometric evaluation of immunoblot analysis for p94 in normal and 4-day stimulated EDL muscles from rabbit. Values (means ±SD, n = 4) in the stimulated muscles are expressed as percentages of the contralateral control muscles (n = 4). Inset: comparison of p94 contents in normal fast-twitch tibialis anterior (TA) and slow-twitch soleus (Sol) muscles (means ± SD, n = 4). Data are arbitrary units. The integral densitometric intensities of the p94 bands were corrected to total protein as determined by Ponceau S staining of the blots. **P < 0.01.


It is conceivable that the exchange of myofibrillar protein isoforms during CLFS-induced fiber type transitions depends on both synthesis of newly expressed isoforms and proteolysis of isoforms no longer expressed. Upregulation of the proteasome system has, in fact, been demonstrated in a recent study to occur in low-frequency stimulated rabbit fast-twitch muscle (16). Because myofibrillar proteins cannot be directly degraded by the proteasome (28), an additional proteolytic system is required to initiate myofibrillar protein degradation.

In a recent study on rat muscle undergoing CLFS-induced fiber type transition, we provided evidence that increases in calpain activity represent an early event of the transformation process. Similar to exercise (1), elevated calpain activity in low-frequency stimulated muscle was shown to be accompanied by changes in intracellular location. This cellular redistribution encompassed a translocation of cytosolic calpain to myofibrillar and microsomal structures. It was more pronounced for μ-calpain than for m-calpain (32). Because CLFS-induced fiber transformation in fast-twitch rat muscle occurs without major necrotic events (4,23) and m-calpain has been reported to be involved in muscle fiber necrosis (30), we assumed that μ-calpain plays a specific role in the transformation process of uninjured, mature fibers. Moreover, we suggested translocation of μ-calpain to represent an important step for controlled proteolysis during muscle fiber transformation (32).

The present study focuses on transforming rabbit muscle where fast-to-slow conversion has been shown to include fiber atrophy, fiber transformation, and replacement of necrotic fibers by newly formed, satellite cell-derived myotubes (13, 14). These three different processes most likely encompass functions of different proteinases. It was necessary, therefore, to correlate proteolytic events to muscle morphology in the present study.

We show that increases in calpain activity of low-frequency stimulated rabbit muscle reach similar levels previously demonstrated in transforming rat muscle (32). As revealed by immunohistochemistry, this increase most likely results from three different events: 1) a rise of μ-calpain predominantly in type IIA fibers, 2) elevated levels of m-calpain in necrotic fibers, and 3) infiltrates of calpain-positive, mononucleated cells. Together, the rise of μ-calpain in type IIA fibers and the previously demonstrated translocation of μ-calpain in rat muscle emphasize its suggested role in the transformation process. Conversely, the role of m-calpain appears to be important during fiber degeneration, as shown by its increase in injured or necrotic fibers. The colocalization of μ- and m-calpains with calpastatin both in morphologically intact and injured fibers is noteworthy and is in line with the suggestion that proteinases and their physiological inhibitors are coordinately expressed (35). It is evident, however, that colocalization of calpains and calpastatin does not provide information on calpain activity because their interaction may be modified, e.g., by phosphorylation (22). Such information might be obtained by single fiber analyses, e.g., by in situ zymography.

In view of the established role of the ubiquitin-proteasome system in muscle catabolism (11), enhanced ubiquitination of necrotic fibers in the stimulated rabbit muscle is conceivable. This finding is also in agreement with increased proteasome levels previously demonstrated in homogenates of low-frequency stimulated rabbit muscle (16). As shown in the present study, however, enhanced ubiquitination is also detectable in a considerable fraction of uninjured fibers, namely in 34% of all type IIA fibers. The detection of ubiquitin protein conjugates not only in necrotic but also in “morphologically intact” fibers could indicate that ubiquitination is involved also in fiber transformation. However, microlesions undetectable by light microscopic inspection cannot be excluded for these fibers. Furthermore, the failure to detect enhanced ubiquitination in low-frequency stimulated rat muscle reinforces the suggestion that activation of the ubiquitin-proteasome pathway in rabbit muscle relates to muscle damage and atrophy (11).

The observation that a major fraction of the type IIA fibers in rabbit muscle displays elevations in μ-calpain and, to a lesser degree, also in ubiquitin protein conjugates, indicates that these fibers are especially affected by the CLFS-induced increase in contractile activity. This is in line with previous measurements of force production and electromyographic activity in stimulated rabbit muscle in which a steep decay in force output a few minutes after the onset of CLFS was attributed to a transient refractoriness of a major fiber population (7). According to phosphocreatine analyses on single MHC-classified fibers from low-frequency stimulated muscles, noncontracting fibers were identified as type IID, whereas the majority of the contracting fibers was identified as type IIA (3). These observations must be taken into account when interpreting the changes in proteolytic profiles of the present study. It appears likely that contracting type IIA fibers are more affected by CLFS than noncontracting type IID fibers and, therefore, respond with an enhanced proteolytic activity. Enhanced proteolysis may not only relate to fast-to-slow transitions in myofibrillar protein isoforms but may also apply to other cellular elements, e.g., the enzyme apparatus of energy metabolism (20) and proteins involved in Ca2+-dynamics (7, 15).

To our knowledge, reduced expression levels of the skeletal muscle-specific p94 have been observed only under pathological conditions, namely in limb girdle muscular dystrophy (24). The rapid disappearance of p94 in low-frequency stimulated rabbit muscle may be a direct effect of enhanced neuromuscular activity. It might be related to its specific binding to titin (29). Its release could destabilize titin or result from titin cleavage during the induced fast-to-slow transition.

In summary, we show that CLFS of rabbit fast-twitch muscle enhances calpain activity and ubiquitination as early events of fast-to-slow fiber transitions. As judged by immunohistochemistry, m-calpain is prominent mainly in necrotic fibers, whereas μ-calpain is elevated in intact fibers. According to MHC-based fiber typing, increases in μ-calpain and enhanced ubiquitination predominantly encompass type IIA fibers.


This study was supported by Deutsche Forschungsgemeinschaft Grant Pe 62/27.


  • Address for reprint requests and other correspondence: D. Pette, Department of Biology, Univ. of Konstanz, D-78457 Konstanz, Germany (E-mail: dirk.pette{at}

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract