The influence of water/air cooling on collateral tissue damage using a diode laser with an innovative pulse design (micropulsed mode)-an in vitro study.

Since the diode laser is a good compromise for the daily use in dental offices, finding usage in numerous dental indications (e.g., surgery, periodontics, and endodontics), the minimization of the collateral damage in laser surgery is important to improve the therapeutical outcome. The aim of this study was to investigate the effect of water/air cooling on the collateral thermal soft tissue damage of 980-nm diode laser incisions. A total of 36 mechanically executed laser cuts in pork liver were made with a 980-nm diode laser in micropulsed mode with three different settings of water/air cooling and examined by histological assessment to determine the area and size of carbonization, necrosis, and reversible tissue damage as well as incision depth and width. In our study, clearly the incision depth increased significantly under water/air cooling (270.9 versus 502.3 µm-test group 3) without significant changes of incision width. In test group 2, the total area of damage was significantly smaller than in the control group (in this group, the incision depth increases by 65 %). In test group 3, the total area of damage was significantly higher (incision depth increased by 85 %), but the bigger part of it represented a reversible tissue alteration leaving the amount of irreversible damage almost the same as in the control group. This first pilot study clearly shows that water/air cooling in vitro has an effect on collateral tissue damage. Further studies will have to verify, if the reduced collateral damage we have proved in this study can lead to accelerated wound healing. Reduction of collateral thermal damage after diode laser incisions is clinically relevant for promoted wound healing.

Reduction of collateral thermal impact of diode laser irradiation on soft tissue due to modified application parameters.

The aim of this study was to investigate the effect of different working modes (pulsed and micropulsed) and power settings of a standardized 980-nm diode laser on collateral thermal soft-tissue damage. A total of 108 bovine liver samples were cut with a diode laser at various settings in pulsed and micropulsed mode and histologically assessed to determine the area and depth of carbonization, necrosis and reversible tissue damage, as well as incision depth and width. Incision depth and width and the area and depth of carbonization, necrosis and reversible damage were correlated strongly with cutting speed. The area and depth of reversible damage were correlated with average power. The micropulsed mode produced a smaller zone of carbonization and necrosis and a smaller incision width. Setting the laser parameters in accordance with the absorption characteristics of the tissue reduced collateral thermal tissue damage while maintaining an acceptable cutting ability. Reducing collateral thermal damage from diode laser incisions is clinically relevant for promoting wound healing.

Expression and functional characterization of a pHis-tagged human bradykinin B2 receptor in COS-7 cells.

A polyHis-tagged bradykinin (BK) B2 receptor (pHis-BKR) cDNA was constructed and expressed in COS-7 cells. The pHis-BKR is suitable for both immunoprecipitation and immunoblotting with anti-polyHis antibodies and can be easily purified using Ni-NTA columns. Immunochemical detection revealed a molecular mass of approximately 66 kDa. The pHis-BKR is capable of mediating BK-induced stimulation of inositol phosphate formation as well as of mitogen-activated protein kinase (MAPK) activity. Compared with the wild-type receptor (WT-BKR) the tagged receptor showed a slightly enhanced affinity towards BK but a reduced expression level. Despite these modified pharmacological properties the pHis-tagged BKR may be a useful tool for studying BKR modifications and signaling.

Effect of site-directed mutagenesis of the arginine residues 509 and 748 on mouse band 3 protein-mediated anion transport.

Using site-directed mutagenesis, the arginine residues 509 and 748 in mouse band 3 protein were substituted by Lys, Thr, and Cys, or by Lys and Gln, respectively. After expression in Xenopus oocytes of the cRNAs encoding wild type band 3 or any one of the band 3 mutants, chloride equilibrium exchange was measured. When the flux measurements were performed two to three days after microinjection of the cRNAs, in contrast to the wild type, neither one of the mutants was able to accomplish transport, with the possible exception of the mutants R509K and R748K both of which showed some transport activity of doubtful significance. Immunoprecipitates revealed that the Arg 748 mutants were expressed similar to the wild type band 3 while no expression of the Arg 509 mutants could be detected. When the flux measurements were performed only 3 h after microinjection of the cRNAs, transport activity was observed in the oocytes that had received cRNAs encoding wild type band 3. In some oocytes of a population, a very slight transport activity was brought about by cRNA encoding Arg 509 mutants. No transport activity could be detected after injection of the Arg 748 mutant. Immunoprecipitation demonstrated the successful biosynthesis of wild type band 3 and of both the Arg 509 and the Arg 748 mutants. The experiments suggest that mutation of Arg 748 leads to biosynthesis of an inactive form of the band 3 protein, while that of Arg 509 results in expression of an abnormally folded, possibly functionally more or less intact form, which is proteolytically degraded within less than one day.

Three different actions of phenylglyoxal on band 3 protein-mediated anion transport across the red blood cell membrane.

Phenylglyoxalation of the red blood cell membrane leads to three superimposed effects on band 3 protein-mediated anion equilibrium exchange as measured by means of radiosulfate: (1) a shift of the curve relating transport activity to pH towards lower pH values, possibly in combination with an increase of the maximal transport activity. This is accompanied by effect (2), the abolishment of a chloride-stimulated component of anion transport seen at low pH values. Effect (3) consists of inhibition of anion equilibrium exchange. Effect (1) prevails when phenylglyoxalation is performed at low concentrations of PG and low pH, while effect (3) predominates when exposure to PG is executed at high pH and high concentration of PG. Effect (1) is associated with a decrease of the Ki values for inhibition and binding of the reversibly acting stilbene disulfonates DNDS and DBDS. The inhibition observed as a consequence of effect (3) is linearly related to a decrease of the capacity of band 3 to combine with the stilbene disulfonate DBDS. The results are interpreted on the assumption that PG is capable of reacting with two or possibly three distinct binding sites in band 3. Reaction with one of them leads to effect (1) and, perhaps, to effect (2); reaction with the other to effect (3). The latter is possibly due to modification of Arg 730, which is homologous to Arg 748 in mouse band 3. Site-directed mutagenesis of this arginine residue showed that it is required for band 3-mediated anion transport.

Transport-related conformational states of the band 3 protein: probing with 1-fluoro-2,4-dinitrobenzene.

The present article provides experimental evidence for previous claims, that Lys 539, without being directly involved in anion binding or translocation, is allosterically linked to the anion binding sites of the band 3 protein and to some other, as yet unidentified amino acid residue. The evidence is based on a detailed study of the kinetics of inhibition of sulphate equilibrium exchange by 1-fluoro-2,4-dinitrobenzene (N2ph-F). It is shown that the mutation of Lys 558 in mouse band 3, which is homologous to Lys 539 in human band 3, renders the transport protein insusceptible to inhibition by N2pH-F, confirming that it is the modification of this residue which results in the inhibition of band 3-mediated transport. The investigation of the kinetics of the modification of human band 3 revealed that the modification is not preceded by non-covalent N2ph-F binding and hence governed by the structure of the native protein near Lys 539. In chloride-containing media, the rate constant of dinitrophenylation of Lys 539 is about 15 times higher than in sulphate-containing media. This suggests that the chemical nature of the anion species bound to band 3 determines whether Lys 539 exists in a buried or exposed state and hence represents a reporter group which characterizes the functional state of the transport protein. The parameter values describing the effects of anion binding on the interactions between Lys 539 and an allosterically linked, unidentified amino acid residue were determined by means of a mathematical model which permitted the quantitative evaluation of the data.

Inhibition of mouse erythroid band 3-mediated chloride transport by site-directed mutagenesis of histidine residues and its reversal by second site mutation of Lys 558, the locus of covalent H2DIDS binding.

Substitution by site-directed mutagenesis of any one of the histidine residues H721, H837, and H852 by glutamine, or of H752 by serine, inhibits Cl- flux mediated by band 3 expressed in Xenopus oocytes. Mutation of Lys 558 (K558N), the site of covalent binding of H2DIDS (4,4'-diisothiocyanostilbene-2,2'-disulfonate) in the outer membrane surface, in combination with any one of the His/Gln mutations leads to partial (H721Q; H837Q) or complete (H852Q) restoration of Cl- flux. In contrast, inhibition of Cl- flux by mutation of proline or lysine residues in the vicinity of His 837 at the inner membrane surface cannot be reversed by the second-site mutation K558N, indicating specificity of interaction between Lys 558 and His 837. The histidine-specific reagent diethyl pyrocarbonate (DEPC) is known to inhibit band 3-mediated anion exchange in red blood cells [Izuhara, K., Okubo, K., & Hamasaki, N. (1989) Biochemistry 28, 4725-4728]. It was also found to inhibit transport after expression in the oocyte of wild-type band 3, of the double mutants of the histidines listed above, and of the single mutant H752S. The effects on the wild type and the double mutants were indistinguishable, while the mutant H752S exhibited a considerably reduced sensitivity to inhibition, suggesting that His 752 is the most prominent site of action of DEPC. According to a hydrophobicity plot of band 3 and further independent evidence, Lys 558, the mutated histidines, and Glu 699, the mutation of which was also found to inhibit Cl- flux [Müller-Berger, S., Karbach, D., Kang, D., Aranibar, N., Wood, P. G., Rüterjans, H., & Passow, H. (1995) Biochemistry 34, 9325-9332], are most likely located in five different transmembrane helices. The interactions between Lys 558 and the various histidines suggest that these helices reside in close proximity. Together with the helix carrying Glu 699, they could form an access channel lined with an array of alternating histidine and glutamate residues. Together with a chloride ion bridging the gap between His 852 and His 837, they could have the potential to form, at low pH, a transmembrane chain of hydrogen bonds. The possible functional significance of such channel is discussed.

Roles of histidine 752 and glutamate 699 in the pH dependence of mouse band 3 protein-mediated anion transport.

In the accompanying paper we have shown that four different histidine residues are involved in the maintenance of mouse band 3 in a state in which it is able to execute its anion transport function. Here we focus on the functional significance of His 752 and demonstrate that this residue, together with Glu 699, plays a key role in the control of pH dependence of Cl- transport. Mouse band 3-encoding cRNA was expressed in Xenopus oocytes, and band 3-mediated Cl- transport was measured at zero membrane potential over the pH range 6.0-9.2. Transport decreased with increasing H+ concentration and was governed by a single pK of 5.8. After correction for temperature differences, this result agrees well with measurements in erythrocyte ghosts of Cl- flux by Funder and Wieth [Funder, J., & Wieth, J. O. (1976) J. Physiol. 262, 679-698] and our own determinations by 35Cl NMR spectroscopy of Cl- exchange between the substrate binding site and the medium. After mutation of either Glu 699 to Asp or of His 752 to Ser, the maximal rate of transport is reduced and the rate of anion exchange is now governed by a single pK of about 6.8-6.9. This suggests that the formation of a hydrogen bond between His 752 and Glu 699 is essential for the decrease of band 3-mediated Cl- transport at low pH. We suggest that in the wild type band 3 both the decrease of the chloride exchange between the medium and the substrate binding site and the inhibition of chloride translocation across the membrane are dominated by a common rate-limiting step and that this step involves hydrogen bond formation between Glu 699 and His 752.

Calcium pump kinetics determined in single erythrocyte ghosts by microphotolysis and confocal imaging.

The activity of the plasma membrane calcium pump was measured in single cells. Human red blood cell ghosts were loaded with a fluorescent calcium indicator and either caged calcium and ATP (protocol A) or caged ATP and calcium (protocol B). In a suitably modified laser scanning microscope either calcium or ATP were released by a short UV light pulse. The time-dependent fluorescence intensity of the calcium indicator was then followed in single ghosts by repetitive confocal imaging. The fluorescence intensity was converted into calcium concentration, which in turn was used to derive the kinetic parameters of the calcium pump, the Michaelis-Menten constant Km, and the maximal transport rate vmax. Km and vmax values derived in this manner were 24 +/- 14 microM and 1.0 +/- 0.6 microM/(ghost s) for protocol A, and 4 +/- 3 microM and 1.0 +/- 0.6 microM/(ghost s) for protocol B, respectively. The difference between A and B is presumably caused by calmodulin, which is inactive in the experiments with protocol A. The possibilities to extend the new method to living nucleus-containing cells transiently transfected with mutants of the plasma membrane calcium pump are discussed.

The influence of two anion-transport inhibitors, 4,4'-diisothiocyanatodihydrostilbene-2,2'-disulfonate and 4,4'-dibenzoylstilbene-2,2'-disulfonate, on the self-association of erythrocyte band 3 protein.

4,4'-Diisothiocyanatodihydrostilbene-2,2'-disulfonate and 4,4'-dibenzoylstilbene-2,2'-disulfonate potently inhibit the erythrocyte anion transporter. These inhibitors act by binding, with a 1:1 stoichiometry, to the band 3 transport protein. We have studied, by sedimentation equilibrium analysis in an analytical ultracentrifuge, the effect of the two closely related stilbenedisulfonates on the state of association of band 3 in the nonionic detergent nonaethyleneglycol lauryl ether. It was found that covalent binding of 4,4'-diisothiocyanatodihydrostilbene-2,2'-disulfonate to band 3 did not significantly disturb the monomer/dimer/tetramer association equilibrium shown by the unliganded protein. An entirely different result was obtained after addition of 4,4'-dibenzoylstilbene-2,2'-disulfonate to the protein, at both low and high chloride concentrations. The amount of band 3 dimer in the samples increased with increasing inhibitor concentration c1, and for c1 > or = 15 microM virtually all of the protein was present as dimer. After removal of the inhibitor (by gel filtration or dialysis), the original monomer/dimer/tetramer distribution of the band 3 protein was restored. Our data show that the (noncovalent) binding of 4,4'-dibenzoylstilbene-2,2'-disulfonate drastically changes the coupling between band 3 protomers. In addition, a reversible change in the state of association of band 3 induced by ligand binding is demonstrated.

Anion transport function of mouse erythroid band 3 protein (AE1) does not require acylation of cysteine residue 861.

Cys-861 of mouse band 3 is equivalent to Cys-843 of human band 3, the only acylated cysteine residue in the anion exchanger AE1 of the red blood cell (Hamasaki et al. (1992) Progress Cell Res. 2, 65-71). Mutation of Cys-861 to serine or methionine caused no significant changes of band 3-mediated anion exchange as measured after expression of the appropriate cRNAs in Xenopus oocytes. Susceptibility to inhibition of transport by 4,4'-dinitrostilbene-2,2'-disulfonate and PCMBS was not affected. We conclude that palmitoylation is not an absolute requirement for the successful execution the anion transport function by the hydrophobic domain of band 3 in the plasma membrane.

Role of Lys 558 and Lys 869 in substrate and inhibitor binding to the murine band 3 protein: a study of the effects of site-directed mutagenesis of the band 3 protein expressed in the oocytes of Xenopus laevis.

The effect of mutation of either Lys 558 or Lys 869 or both on mouse erythroid band 3 protein (AE1)-mediated 36Cl- efflux and its inhibition by pyridoxal 5-phosphate (P5-P), DNDS and H2DIDS were studied. Regardless of the mutation, band 3 was always capable of executing Cl- self-exchange. P5-P (5 mM, pH 7.6) produced irreversible inhibition in the wild type (KK) and in the mutant in which Lys 558 (NK) or Lys 869 (KM) had been replaced by asparagine (N) or methionine (M), respectively. However, when both residues were replaced, mutant (NM), irreversible inhibition could no longer be achieved. This shows that P5-P is capable of producing inhibition with either one of the lysine residues, 558 or 869. Inhibition by DNDS changed dramatically upon mutation. The Ki app increased from 6.0 microM in the wild type (KK) to 23 microM in the mutant NK, to 73 microM in the mutant KM and to 474 microM in the double mutant NM. The Km value for activation of the transport system by varying the substrate concentration by isosmotic substitution of Cl- with SO4(2-) decreased from 42 mM in the wild type (KK) to 11.3 mM in the mutant NM. The results show that both Lys 558 and Lys 869 are involved in the maintenance of the structure of the overlapping binding sites for stilbene disulfonates and the substrate Cl-. In the double mutant NM, H2DIDS is no longer able to produce irreversible inhibition at pH 7.6. This is evidently related to the replacement of Lys 558 (pK 8.2) by Asn 558 in this mutant (see Bartel, D., Lepke, S., Layh-Schmitt, G., Legrum, B., Passow, H., 1989. EMBO J. 8:3601-3609). However, at pH 9.5, some irreversible inhibition could still be observed. This suggests that the other lysine residue (pK 10.8) that is known to be involved in covalent binding with the second isothiocyanate group of H2DIDS is still present, and hence, not identical to Lys 869, which had been substituted by a methionine residue. However, this result remains inconclusive since after mutagenesis, the H2DIDS may produce inhibition at a site that is not normally involved in H2DIDS binding.

Mediation of inorganic anion transport by the hydrophobic domain of mouse erythroid band 3 protein expressed in oocytes of Xenopus laevis.

A cDNA clone of the mouse erythroid band 3 protein encoding the 556 amino acid residues of the hydrophobic domain from Thr-374 to the C-terminal Val-929 is shown by immunoprecipitation to be expressed in Xenopus oocytes. Measurements of 36Cl- efflux indicate that the translation product mediates Cl- transport, which is inhibitable reversibly by DNDS or H2DIDS, specific inhibitors of band 3-mediated transport. The apparent KI values are 3.6 microM and 0.094 microM, respectively, and hence similar to those found in the wild type band 3-mediated anion transport. The rapid reversible inhibition by H2DIDS slowly changes to irreversible inhibition. The rate of change increases with increasing pH, again similar as to the wild-type band 3. It is concluded that the hydrophobic domain of band 3 is capable of executing anion transport essentially similar to the full-length band 3, although minor differences with respect to transport and inhibition kinetics cannot be ruled out.

pH-dependence of inhibition by H2DIDS of mouse erythroid band 3-mediated Cl- transport in Xenopus oocytes. The effect of oligonucleotide-directed replacement of Lys-558 by an Asn residue.

The rapid reversible inhibition of band 3-mediated inorganic anion transport by 4,4'-diisothiocyanodihydrostilbene-2,2'-disfulfonate (H2DIDS) turns slowly into irreversible inhibition. This is due to covalent bond formation of the two isothiocyanate groups of the inhibitor with two lysine residues on band 3, called Lys a and Lys b. In the red cell membrane, the pK value of Lys a is about 2.5 pK units lower than the pK value of Lys b. Hence the susceptibility of Lys a to irreversible modification by H2DIDS far exceeds the susceptibility to Lys b. In the present paper, we have expressed in Xenopus oocytes cRNA's derived from cDNA clones encoding wild-type mouse band 3 and mouse band 3 in which Lys a (Lys-558) had been replaced by an Asn residue by oligonucleotide-directed mutagenesis. In accord with previous findings, in the oocytes both wild-type and mutated band 3 mediate Cl- exchange. After determining the uninhibited exchange rate the oocytes were exposed for a fixed length of time to H2DIDS at a concentration (20 microM) which saturates all H2DIDS binding sites with reversibly bound H2DIDS (KI = 0.3 microM and 1.1 microM, respectively, for wild-type and mutant). Exposure was terminated by washing with a medium in which H2DIDS was replaced by bovine serum albumin to remove free and reversibly bound H2DIDS from the extracellular phase. Subsequent measurements of Cl- efflux yielded a measure for the irreversible inhibition that persisted. Since the transition from reversible to irreversible H2DIDS binding was found to follow first-order kinetics it was possible to calculate rate constants. From the pH dependence of the rate constants, pK values were calculated. These calculations could be made since in the wild-type, in which Lys a and Lys b are present, the exposure to H2DIDS could be confined to a pH range in which little if any covalent binding to Lys b takes place. The data could be represented by a single pK value of 8.3. In the mutant, Lys a is missing. Hence, covalent reaction can only take place with Lys b. Measurements over the appropriate pH range could be described by a single pK of 10.8. These values are 0.8-0.9 pK units higher than those previously obtained in experiments with band 3 in the red cell membrane (Kampmann et al. (1982) J. Membr. Biol. 70, 199-216).(ABSTRACT TRUNCATED AT 250 WORDS)

Downregulation of surface sodium pumps by endocytosis during meiotic maturation of Xenopus laevis oocytes.

During meiotic maturation, plasma membranes of Xenopus laevis oocytes completely lose the capacity to transport Na and K and to bind ouabain. To explore whether the downregulation might be due to an internalization of the sodium pump molecules, the intracellular binding of ouabain was determined. Selective permeabilization of the plasma membrane of mature oocytes (eggs) by digitonin almost failed to disclose ouabain binding sites. However, when the eggs were additionally treated with 0.02% sodium dodecyl sulfate (SDS) to permeabilize inner membranes, all sodium pumps present before maturation were recovered. Phosphorylation by [gamma-32P]ATP combined with SDS-polyacrylamide gel electrophoresis (PAGE) and autoradiography showed that sodium pumps were greatly reduced in isolated plasma membranes of eggs. According to sucrose gradient fractionation, maturation induced a shift of sodium pumps from the plasma membrane fraction to membranes of lower buoyant density with a protein composition different from that of the plasma membrane. Endocytosed sodium pumps identified on the sucrose gradient from [3H]ouabain bound to the cell surface before maturation could be phosphorylated with inorganic [32P]phosphate. The findings suggest that downregulation of sodium pumps during maturation is brought about by translocation of surface sodium pumps to an intracellular compartment, presumably endosomes. This contrasts the mechanism of downregulation of Na-dependent cotransport systems, the activities of which are reduced as a consequence of a maturation-induced depolarization of the membrane without a removal of the corresponding transporter from the plasma membrane.

Anion transport in oocytes of Xenopus laevis induced by expression of mouse erythroid band 3 protein--encoding cRNA and of a cRNA derivative obtained by site-directed mutagenesis at the stilbene disulfonate binding site.

A vector was constructed containing a cDNA for mouse band 3 obtained from Demuth et al. (1986, EMBO J., 5, 1205-1214), a synthetic linker (containing 5'-non-translated region, start codon and a coding region for the first 12 N-terminal amino acids), and RNA polymerase promoters suitable for in vitro transcription of cRNA. After injection of the cRNA into the cytoplasm of Xenopus oocytes and incubation for 16 h, expression of mouse band 3 was demonstrated by immunoprecipitation, immunohistochemical methods and influx or efflux measurements with 36Cl-. Antisense cRNA inhibits the expression. Lysines 558 and 561 were replaced by asparagines using oligonucleotide-directed mutagenesis. Like the original band 3, the mutant shows stilbene disulfonate-inhibitable anion exchange. However, in contrast to the original band 3, inhibition by 4,4'-diisothiocyano dihydrostilbene-2,2'-disulfonate (H2DIDS) is no longer irreversible. This indicates that thiourea bond formation between H2DIDS and band 3 involves one of the two modified lysine residues. It also shows that the two lysine residues are not essential for the execution of the anion transport function of band 3. The results described suggest that the cDNA clone of Demuth et al. (1986) encodes a protein with properties that are representative for the properties of the bulk of the band 3 protein in the plasma membrane of the red cell of the mouse.

Identification by site-directed mutagenesis of Lys-558 as the covalent attachment site of H2DIDS in the mouse erythroid band 3 protein.

After functional expression of mouse erythroid band 3 by cRNA microinjection into Xenopus oocytes, 36Cl- efflux is irreversibly inhibited by H2DIDS. When a cRNA is injected that is derived from a cDNA in which the nucleotides encoding for lysine-558 were replaced by nucleotides encoding for asparagine, transport and inhibition of transport by H2DIDS still occur. However, when measured under conditions where no intramolecular crosslinking takes place the inhibition by H2DIDS is no longer irreversible. This indicates that thiourea bond formation between H2DIDS and band 3 takes place at Lys-558.

The effects of dansylation on the pH dependence of SO4(2-) and Cl- equilibrium exchange and on the H+/SO4(2-) cotransport across the red blood cell membrane.

Treatment of the erythrocyte membrane with dansyl chloride leads to the following effects: (i) SO4(2-) transport is enhanced, Cl- transport is reduced. At maximal acceleration of sulfate exchange, Cl- exchange is only partially inhibited. The two effects are lineary related suggesting that the Cl- and SO4(2-) transporting forms of band 3 are derived from the same pool. (ii) The maximum of the pH dependence of SO4(2-) equilibrium exchange as measured at low sulfate concentrations is replaced by a plateau. It now resembles the pH dependence of Cl- exchange in untreated red cells. The pH dependence of SO4(2-) equilibrium exchange as measured at high sulfate concentrations is virtually unchanged after dansylation. The pH dependence of the partially inhibited Cl- equilibrium exchange across the dansylated membrane as measured at high chloride concentrations remains similar as in the untreated red cells but is somewhat less pronounced. (iii) SO4(2-)/H+ cotransport remains essentially unaltered after modification by dansyl chloride. The effects of dansylation are discussed in terms of a model similar to the titratable carrier model originally proposed by Gunn (Gunn, R.B. (1972) in Oxygen Affinity of Hemoglobin and Red Cell Acid Base Status (Rorth, M. and Astrup, P., eds.), pp. 823-827, Munksgaard, Copenhagen).