Kinetic studies of iron deposition catalyzed by recombinant human liver heavy and light ferritins and Azotobacter vinelandii bacterioferritin using O2 and H2O2 as oxidants

The discrepancy between predicted and measured H(2)O(2) formation during iron deposition with recombinant heavy human liver ferritin (rHF) was attributed to reaction with the iron protein complex [Biochemistry 40 (2001) 10832-10838]. This proposal was examined by stopped-flow kinetic studies and analysis for H(2)O(2) production using (1) rHF, and Azotobacter vinelandii bacterial ferritin (AvBF), each containing 24 identical subunits with ferroxidase centers; (2) site-altered rHF mutants with functional and dysfunctional ferroxidase centers; and (3) recombinant human liver light ferritin (rLF), containing no ferroxidase center. For rHF, nearly identical pseudo-first-order rate constants of 0.18 s(-1) at pH 7.5 were measured for Fe(2+) oxidation by both O(2) and H(2)O(2), but for rLF, the rate with O(2) was 200-fold slower than that for H(2)O(2) (k = 0.22 s(-1)). A Fe(2+)/O(2) stoichiometry near 2.4 was measured for rHF and its site altered forms, suggesting formation of H(2)O(2). Direct measurements revealed no H(2)O(2) free in solution 0.5-10 min after all Fe(2+) was oxidized at pH 6.5 or 7.5. These results are consistent with initial H(2)O(2) formation, which rapidly reacts in a secondary reaction with unidentified solution components. Using measured rate constants for rHF, simulations showed that steady-state H(2)O(2) concentrations peaked at 14 muM at approximately 600 ms and decreased to zero at 10-30 s. rLF did not produce measurable H(2)O(2) but apparently conducted the secondary reaction with H(2)O(2). Fe(2+)/O(2) values of 4.0 were measured for AvBF. Stopped-flow measurements with AvBF showed that both H(2)O(2) and O(2) react at the same rate (k = 0.34 s(-1)), that is faster than the reactions with rHF. Simulations suggest that AvBF reduces O(2) directly to H(2)O without intermediate H(2)O(2) formation.     

Kinetic studies of iron deposition in horse spleen ferritin using H2O2 and O2 as oxidants

The reaction of horse spleen ferritin (HoSF) with Fe2+ at pH 6.5 and 7.5 using O2, H2O2 and 1:1 a mixture of both showed that the iron deposition reaction using H2O2 is approximately 20- to 50-fold faster than the reaction with O2 alone. When H2O2 was added during the iron deposition reaction initiated with O2 as oxidant, Fe2+ was preferentially oxidized by H2O2, consistent with the above kinetic measurements. Both the O2 and H2O2 reactions were well defined from 15 to 40 degrees C from which activation parameters were determined. The iron deposition reaction was also studied using O2 as oxidant in the presence and absence of catalase using both stopped-flow and pumped-flow measurements. The presence of catalase decreased the rate of iron deposition by approximately 1.5-fold, and gave slightly smaller absorbance changes than in its absence. From the rate constants for the O2 (0.044 s(-1)) and H2O2 (0.67 s(-1)) iron-deposition reactions at pH 7.5, simulations of steady-state H2O2 concentrations were computed to be 0.45 microM. This low value and reported Fe2+/O2 values of 2.0-2.5 are consistent with H2O2 rapidly reacting by an alternate but unidentified pathway involving a system component such as the protein shell or the mineral core as previously postulated [Biochemistry 22 (1983) 876; Biochemistry 40 (2001) 10832].     

Redox reactivity of animal apoferritins and apoheteropolymers assembled from recombinant heavy and light human chain ferritins

The redox reactivities of air-oxidized apo horse spleen ferritin (HoSF) and apo rat liver ferritin (RaF) were examined by microcoulometry and reductive optical titrations. Microcoulometry on several independent lots of commercial HoSF revealed two distinct types of redox activity: one requiring 3-4 electrons and one requiring 6-7 electrons for full reduction of the protein shell. ApoRaF required 8-9 electrons to fully reduce the oxidized form. Reductive optical titrations confirmed the microcoulometric reduction stoichiometry and, in addition, showed that the spectra of both oxidized and reduced apoHoSF were distinct and possessed absorbances tailing into the visible region. The redox reactivity of both apoRaF and apoHoSF correlated with their H-subunit composition. Identical microcoulometric and optical experiments were conducted with recombinant apo human liver heavy (rHuHF) and light (rHuLF) ferritins, but neither was redox-active. These results suggest that the redox reactivity of native ferritins is due to their heteropolymeric nature. This was confirmed by mixing various proportions of rHuHF and rHuLF, dissociating the 24-mers into individual subunits with guanidine hydrochloride at pH 3.5, and renaturing to form heteropolymeric 24-mers. Microcoulometric measurements of these apoheteropolymers reassembled in vitro showed that they were redox-active like their native apoheteropolymer counterparts. The redox activity of these apoheteropolymers increased with H-subunit composition, reached a maximum near 12 H- and 12 L-subunits, and then declined to zero with increasing L-subunit composition. The decline in redox reactivity at high L-subunit concentrations indicates that both H- and L-subunits are involved in forming the observed redox centers. Apoheteropolymers formed from rHuLF and W93F (an H-chain mutant) were redox-inactive, suggesting that the conserved tryptophan is necessary for redox center formation.     

Heterogeneity in horse ferritins. A comparative study of surface charge, iron content and kinetics of iron uptake.

Horse ferritins from different organs show heterogeneity on electrofocusing in Ampholine gradients. Both ferritin and apoferritin from liver and spleen could be fractionated with respect to surface charge by serial precipitation with (NH4)2SO4. In the ferritin fractions, increasing iron content parallels increasing isoelectric point. After removal of their iron, those fractions which originally contained most iron accumulated added iron at the fastest rates. When unfractionated ferritins from different organs were compared the average isoelectric point increased in order spleen less than liver less than kidney less than heart. The order of initial rates of iron uptake by the apoferritins was spleen greater than kidney greater than heart and initial average iron contents also followed this order. The relatively low rates of iron accumulation by iron-poor molecules may have been due to structural alteration, to degradation, to activation of the iron-rich molecules or to other factors.     

Evidence that a salt bridge in the light chain contributes to the physical stability difference between heavy and light human ferritins.

Human ferritin, a multimeric iron storage protein, is composed by various proportions of two subunit types: the H- and L-chains. The biological functions of these two genic products have not been clarified, although differences in reactivity with iron have been shown. Starting from the hypothesis that the high stability typical of ferritin is an important property which may be relevant for its iron storage function, we studied ferritin homopolymers of H- and L-chains in different denaturing conditions. In addition we analyzed 13 H-chain variants with alterations in regions conserved within mammalian H-chains. In all the denaturation experiments H-chain ferritin showed lower stability than L-chain ferritin. The difference was greater in guanidine HCl denaturation experiments, where the end products are fully unfolded peptides, than in acidic denaturation experiments, where the end products are peptides with properties analogous to "molten globule." The study on H-chain variants showed: (i) ferritin stability was not affected by alterations of regions exposed to the inner or outer surface of the shell and not involved in intra- or inter-chain interactions; (ii) stability was reduced by alterations of sequences involved in inter-subunit interactions such as the deletion of the N-terminal extension or substitutions along the hydrophobic and hydrophilic channels; (iii) stability was increased by the substitution of 2 amino acids inside the four-helix bundle with those of the homologous L-chain. One of the residues is involved in a salt bridge in the L-chain, and we concluded that the stability difference between H- and L-ferritins is to a large extent due to the stabilizing effect of this salt bridge on the L-subunit fold.     

Formation and movement of Fe(III) in horse spleen, H- and L-recombinant ferritins. A fluorescence study.

Iron oxidation and incorporation into apoferritins of different subunit composition, namely the recombinant H and L homopolymers and the natural horse spleen heteropolymer (10-15% H), have been followed by steady-state and time-resolved fluorescence. After aerobic addition of 100 Fe(II) atoms/polymer, markedly different kinetic profiles are observed. In the rL-homopolymer a slow monotonic fluorescence quenching is observed which reflects binding, slow oxidation at the threefold apoferritin channels, and diffusion into the protein cavity. In the rH-homopolymer a fast fluorescence quenching is followed by a partial, slow recovery. The two processes have been attributed to Fe(II) binding and oxidation at the ferroxidase centers and to Fe(III) released into the cavity, respectively. The fluorescence kinetics of horse spleen apoferritin is dominated by the H chain contribution and resembles that of the H homopolymer. It brings out clearly that the rate of the overall process is limited by the rate at which Fe(III) leaves the ferroxidase centers of the H chains where binding of incoming Fe(II) and its oxidation take place. The data obtained upon stepwise addition of iron and the results of optical absorption measurements confirm this picture. The correspondence between steady-state and time-resolved data is remarkably good; this is manifest when the latter are used to calculate the change in fluorescence intensity as apparent in the steady-state measurements.     

Anaerobic and aerobic oxidation of ferrous iron at neutral pH by chemoheterotrophic nitrate-reducing bacteria

Nine out of ten anaerobic enrichment cultures inoculated with sediment samples from various freshwater, brackish-water, and marine sediments exhibited ferrous iron oxidation in mineral media with nitrate and an organic cosubstrate at pH 7.2 and 30° C. Anaerobic nitrate-dependent ferrous iron oxidation was a biological process. One strain isolated from brackish-water sediment (strain HidR2, a motile, nonsporeforming, gram-negative rod) was chosen for further investigation of ferrous iron oxidation in the presence of acetate as cosubstrate. Strain HidR2 oxidized between 0.7 and 4.9 mM ferrous iron aerobically and anaerobically at pH 7.2 and 30° C in the presence of small amounts of acetate (between 0.2 and 1.1 mM). The strain gained energy for growth from anaerobic ferrous iron oxidation with nitrate, and the ratio of iron oxidized to acetate provided was constant at limiting acetate supply. The ability to oxidize ferrous iron anaerobically with nitrate at approximately pH 7 appears to be a widespread capacity among mesophilic denitrifying bacteria. Since nitrate-dependent iron oxidation closes the iron cycle within the anoxic zone of sediments and aerobic iron oxidation enhances the reoxidation of ferrous to ferric iron in the oxic zone, both processes increase the importance of iron as a transient electron carrier in the turnover of organic matter in natural sediments. Received: 24 April 1997 / Accepted: 22 September 1997     

Ferritin-catalyzed consumption of hydrogen peroxide by amine buffers causes the variable Fe2+ to O2 stoichiometry of iron deposition in horse spleen ferritin

Ferritin catalyzes the oxidation of Fe2+ by O2 to form a reconstituted Fe3+ oxy-hydroxide mineral core, but extensive studies have shown that the Fe2+ to O2 stoichiometry changes with experimental conditions. At Fe2+ to horse spleen ferritin (HoSF) ratios greater than 200, an upper limit of Fe2+ to O2 of 4 is typically measured, indicating O2 is reduced to 2H2O. In contrast, a lower limit of Fe2+ to O2 of approximately 2 is measured at low Fe2+ to HoSF ratios, implicating H2O2 as a product of Fe2+ deposition. Stoichiometric amounts of H2O2 have not been measured, and H2O2 is proposed to react with an unknown system component. Evidence is presented that identifies this component as amine buffers, including 3-N-morpholinopropanesulfonic acid (MOPS), which is widely used in ferritin studies. In the presence of non-amine buffers, the Fe2+ to O2 stoichiometry was approximately 4.0, but at high concentrations of amine buffers (0.10 M) the Fe2+ to O2 stoichiometry is approximately 2.5 for iron loadings of eight to 30 Fe2+ per HoSF. Decreasing the concentration of amine buffer to zero resulted in an Fe2+ to O2 stoichiometry of approximately 4. Direct evidence for amine buffer modification during Fe2+ deposition was obtained by comparing authentic and modified buffers using mass spectrometry, NMR, and thin layer chromatography. Tris(hydroxymethyl)aminomethane, MOPS, and N-methylmorpholine (a MOPS analog) were all rapidly chemically modified during Fe2+ deposition to form N-oxides. Under identical conditions no modification was detected when amine buffer, H2O2, and O2 were combined with Fe2+ or ferritin separately. Thus, a short-lived ferritin intermediate is required for buffer modification by H2O2. Variation of the Fe2+ to O2 stoichiometry versus the Fe2+ to HoSF ratio and the amine buffer concentration are consistent with buffer modification.     

Mechanism of ferritin iron uptake: activity of the H-chain and deletion mapping of the ferro-oxidase site. A study of iron uptake and ferro-oxidase activity of human liver, recombinant H-chain ferritins, and of two H-chain deletion mutants

To study the functional differences between human ferritin H- and L-chains and the role of the protein shell in the formation and growth of the ferritin iron core, we have compared the kinetics of iron oxidation and uptake of ferritin purified from human liver (90% L) and of the H-chain homopolymer overproduced in Escherichia coli (100% H). As a control for iron autocatalytic activity, we analyzed the effect of Fe(III) on the iron uptake reaction. The results show that the H-chain homopolymer has faster rates of iron uptake and iron oxidation than liver ferritin in all the conditions analyzed and that the difference is reduced in the conditions in which iron autocatalysis in high: i.e. at pH 7 and in presence of iron core. We have also analyzed the properties of two engineered H-chains, one lacking the last 22 amino acids at the carboxyl terminus and the other missing the first 13 residues at the amino terminus. These mutant proteins assemble in ferritin-like proteins and maintain the ability to catalyze iron oxidation. The deletion at the carboxyl terminus, however, prevents the formation of a stable iron core. It is concluded that the ferritin H-chain has an iron oxidation site which is separated from the sites of iron transfer and hydrolysis and that either the integrity of the molecule or the presence of the amino acid sequences forming the hydrophobic channel is necessary for iron core formation.     

Anaerobic oxidation of o-xylene, m-xylene, and homologous alkylbenzenes by new types of sulfate-reducing bacteria

Various alkylbenzenes were depleted during growth of an anaerobic, sulfate-reducing enrichment culture with crude oil as the only source of organic substrates. From this culture, two new types of mesophilic, rod-shaped sulfate-reducing bacteria, strains oXyS1 and mXyS1, were isolated with o-xylene and m-xylene, respectively, as organic substrates. Sequence analyses of 16S rRNA genes revealed that the isolates affiliated with known completely oxidizing sulfate-reducing bacteria of the delta subclass of the class Proteobacteria. Strain oXyS1 showed the highest similarities to Desulfobacterium cetonicum and Desulfosarcina variabilis (similarity values, 98.4 and 98.7%, respectively). Strain mXyS1 was less closely related to known species, the closest relative being Desulfococcus multivorans (similarity value, 86.9%). Complete mineralization of o-xylene and m-xylene was demonstrated in quantitative growth experiments. Strain oXyS1 was able to utilize toluene, o-ethyltoluene, benzoate, and o-methylbenzoate in addition to o-xylene. Strain mXyS1 oxidized toluene, m-ethyltoluene, m-isoproyltoluene, benzoate, and m-methylbenzoate in addition to m-xylene. Strain oXyS1 did not utilize m-alkyltoluenes, whereas strain mXyS1 did not utilize o-alkyltoluenes. Like the enrichment culture, both isolates grew anaerobically on crude oil with concomitant reduction of sulfate to sulfide.     

Delayed light production by blue-green algae,red algae,and purple bacteria

1. Blue-green algae, red algae, and purple bacteria all show the emission of delayed light. 2. The action spectra for the production of delayed light by three species of blue-green algae have one broad band with a peak at 620 mµ. 3. The action spectrum for production of delayed light by the red algae has one peak at 550 mµ with a shoulder from 600 to 660 mµ. 4. The emission spectra of the delayed light from both the blue-green and red algae were the same as from the green algae, Chlorella. 5. The action spectra for the production of delayed light by the different species of purple bacteria tested consisted of one or more bands not resolved between 800 and 900 mµ. 6. The emission spectrum of the delayed light from the purple bacteria was largely at wave lengths longer than 900 mµ.     

Stereoselective inhibition of human, mouse, and horse cholinesterases by bambuterol enantiomers

Bambuterol is a chiral carbamate and a selective inhibitor of butyrylcholinesterase (BChE, EC 3.1.1.8). In order to relate bambuterol selectivity and stereoselectivity of BChE and acetylcholinesterase (AChE, EC 3.1.1.7) of different species, we studied the inhibition of human, mouse, and horse BChE, as well as AChE of human and mouse by (R)- and (S)-bambuterol. AChE and BChE of all studied species were progressively inhibited by both bambuterol enantiomers, with a preference for the (R)-bambuterol whose inhibition rate constants were about five times higher than that of (S)-bambuterol. We observed no significant difference between human and mouse in bambuterol enantiomer BChE inhibition. However, (R)-bambuterol inhibited horse BChE about 14 times slower than human and mouse BChE, and the inhibition rate for (S)-bambuterol was about 18 times slower. Although the primary structure of horse BChE differs from the other two species in 15 amino acids, we presumed that differences in inhibition rates could be attributed to threonine at position 69 located close to the peripheral site of BChE. Since BChE inhibition by bambuterol enantiomers was at least 8000 times faster than that of AChE, both bambuterol enantiomers proved to be selective BChE inhibitors, as was previously shown for racemate.     

In situ phosphorylation of human platelet myosin heavy and light chains by protein kinase C

Treatment of human platelets with 162 nM 12-O-tetradecanoylphorbol-13-acetate (TPA) resulted in phosphorylation of a number of peptides, including myosin heavy chain and the 20-kDa myosin light chain. The site phosphorylated on the myosin heavy chain was localized by two-dimensional peptide mapping to a serine residue(s) in a single major tryptic phosphopeptide. This phosphopeptide co-migrated with a tryptic peptide that was produced following in vitro phosphorylation of platelet myosin heavy chain using protein kinase C. The sites phosphorylated in the 20-kDa myosin light chain in intact cells were analyzed by two-dimensional mapping of tryptic peptides and found to correspond to Ser1 and Ser2 in the turkey gizzard myosin light chain. In vitro phosphorylation of purified human platelet myosin by protein kinase C showed that in addition to Ser1 and Ser2, a third site corresponding to Thr9 in turkey gizzard myosin light chain is also phosphorylated. The phosphorylatable myosin light chains from human platelets were found to consist of two major isoforms present in approximately equal amounts, but differing in their molecular weights and isoelectric points. A third, minor isoform was also visualized by two-dimensional gel electrophoresis. Following treatment with TPA, both the mono- and diphosphorylated forms of each isoform could be visualized, and the sites of phosphorylation were identified. The phosphate content rose from negligible amounts found prior to treatment with TPA to 1.2 mol of phosphate/mol of myosin light chain and 0.7 mol of phosphate/mol of myosin heavy chain following treatment. These results suggest that TPA mediates phosphorylation of both myosin light and heavy chains in intact platelets by activation of protein kinase C.     

Suboxic deposition of ferric iron by bacteria in opposing gradients of Fe(II) and oxygen at circumneutral pH

The influence of lithotrophic Fe(II)-oxidizing bacteria on patterns of ferric oxide deposition in opposing gradients of Fe(II) and O(2) was examined at submillimeter resolution by use of an O(2) microelectrode and diffusion microprobes for iron. In cultures inoculated with lithotrophic Fe(II)-oxidizing bacteria, the majority of Fe(III) deposition occurred below the depth of O(2) penetration. In contrast, Fe(III) deposition in abiotic control cultures occurred entirely within the aerobic zone. The diffusion microprobes revealed the formation of soluble or colloidal Fe(III) compounds during biological Fe(II) oxidation. The presence of mobile Fe(III) in diffusion probes from live cultures was verified by washing the probes in anoxic water, which removed ca. 70% of the Fe(III) content of probes from live cultures but did not alter the Fe(III) content of probes from abiotic controls. Measurements of the amount of Fe(III) oxide deposited in the medium versus the probes indicated that ca. 90% of the Fe(III) deposited in live cultures was formed biologically. Our findings show that bacterial Fe(II) oxidation is likely to generate reactive Fe(III) compounds that can be immediately available for use as electron acceptors for anaerobic respiration and that biological Fe(II) oxidation may thereby promote rapid microscale Fe redox cycling at aerobic-anaerobic interfaces.     

Locations of disulfide bonds and free cysteines in the heavy and light chains of recombinant human factor VIII (antihemophilic factor A).

The locations of disulfide bonds and free cysteines in the heavy and light chains of recombinant human factor VIII were determined by sequence analysis of fragments produced by chemical and enzymatic digestions. The A1 and A2 domains of the heavy chain and the A3 domain of the light chain contain one free cysteine and two disulfide bonds, whereas the C1 and C2 domains of the light chain have one disulfide bond and no free cysteine. The positions of these disulfide bonds are conserved in factor V and ceruloplasmin except that the second disulfide bond in the A3 domain is missing in both factor V and ceruloplasmin. The positions of the three free cysteines of factor VIII are the same as three of the four cysteines present in ceruloplasmin. However, the positions of the free cysteines in factor VIII and ceruloplasmin are not conserved in factor V.     

Rate of iron transfer through the horse spleen ferritin shell determined by the rate of formation of Prussian Blue and Fe-desferrioxamine within the ferritin cavity

Iron (2+ and 3+) is believed to transfer through the three-fold channels in the ferritin shell during iron deposition and release in animal ferritins. However, the rate of iron transit in and out through these channels has not been reported. The recent synthesis of [Fe(CN)6]3-, Prussian Blue (PB) and desferrioxamine (DES) all trapped within the horse spleen ferritin (HoSF) interior makes these measurements feasible. We report the rate of Fe2+ penetrating into the ferritin interior by adding external Fe2+ to [Fe(CN)6]3- encapsulated in the HoSF interior and measuring the rate of formation of the resulting encapsulated PB. The rate at which Fe2+ reacts with [Fe(CN)6]3- in the HoSF interior is much slower than the formation of free PB in solution and is proceeded by a lag period. We assume this lag period and the difference in rate represent the transfer of Fe2+ through the HoSF protein shell. The calculated diffusion coefficient, D approximately 5.8x10(-20) m2/s corresponds to the measured lag time of 10-20 s before PB forms within the HoSF interior. The activation energy for Fe2+ transfer from the outside solution through the protein shell was determined to be 52.9 kJ/mol by conducting the reactions at 10 approximately 40 degrees C. The reaction of Fe3+ with encapsulated [Fe(CN)6]4- also readily forms PB in the HoSF interior, but the rate is faster than the corresponding Fe2+ reaction. The rate for Fe3+ transfer through the ferritin shell was confirmed by measuring the rate of the formation of Fe-DES inside HoSF and an activation energy of 58.4 kJ/mol was determined. An attempt was made to determine the rate of iron (2+ and 3+) transit out from the ferritin interior by adding excess bipyridine or DES to PB trapped within the HoSF interior. However, the reactions are slow and occur at almost identical rates for free and HoSF-encapsulated PB, indicating that the transfer of iron from the interior through the protein shell is faster than the rate-limiting step of PB dissociation. The method described in this work presents a novel way of determining the rate of transfer of iron and possibly other small molecules through the ferritin shell.     

Hemolysis of human erythrocytes by hypochlorous acid is modulated by amino acids,antioxidants, oxidants,membrane-perforating agents and by divalent metals

The optimal conditions under which hypochlorous acid (NaOCl) either hemolyzes human RBC or kills monkey kidney epithelial cells (BGM) in culture had been investigated. While in Hank's balanced salt solution (HBSS), micromolar amounts of NaOCl caused full hemolysis and also killed BGM cells, in D-MEM or RPMI media rich in amino acids, 25-40 mM of hypochlorite were needed to induce cell injury. Cells exposed to high amounts of NaOCl became highly refractory to strong detergents. Hemolysis by NaOCl was strongly inhibited by a large variety of antioxidants. RBC treated by subtoxic concentrations either of peroxide, peroxyl radical, NO, cholesterol, PLA2, PLC as well as by N2, argon or by mixture of CO2 (10%) and O2 (90%) became much more susceptible to lysis by NaOCl. On the other hand, while RBC treated by Fe2+, Co2+, and V2+ and to a lesser extent with Cu2+ became highly resistant to NaOCl hemolysis presumably due to NaOCl decomposition, no such effect was found either with Co2+ or by Mn2+. RBC treated by azide to destroy catalase and then incubated with peroxide and with NaOCl failed to undergo hemolysis due to the ability of peroxide to decompose NaOCl. The inhibitory effects of the divalent metals on NaOCl-induced hemolysis were also substantiated by measuring the decrease in pH and by cyclic voltammetry. The findings that like peroxide, NaOCl also synergizes with membrane-perforating agents and with a protease to kill epithelial cells further implicate such "cocktails" in cell injury in inflammatory conditions. Taken together, because of the capacity of many agents to scavenge NaOCl, tissue damage by NaOCl-generated neutrophils can take place primarily if activated neutrophils closely adhere to target cells to avoid the scavenging effects of amino acids and of antioxidants. Therefore, the significance of the data which had tested the cytotoxic effects of NaOCl using cells suspended only in salt solutions, should be reconsidered.