This study describes the depth distribution of humic-acid-reducing and iron-reducing bacteria in a freshwater sediment, and compares it to the general redox characteristics of humic acids and iron extracted from the corresponding sediment layers. The populations of humic-acid-reducing bacteria were found to be 10–100 times higher than the populations of iron-reducing bacteria in the respective sediment layers.
4.1Microbial humic acid reduction
It is reasonable to assume that bacteria capable of oxidizing either lactate or acetate with humic acids as electron acceptor represent a subset of the total humic-acid-reducing community in the sediment. In a previous study, Coates et al.  have found that all humic-acid-reducing bacteria isolated from various sediments with acetate as electron donor and AQDS as electron acceptor were also able to reduce poorly crystalline ferric iron hydroxide . At first glance, these findings seem to disagree with the results of the present study, where the number of humic-acid-reducing bacteria oxidizing lactate or acetate was much higher than that of bacteria reducing poorly crystalline ferric iron hydroxide with the same electron donors in all sediment layers. However, it must be considered that also pure cultures of fermenting and halorespiring bacteria, which are unable to reduce ferric iron hydroxide directly, are able to use humic acids as electron acceptor [9,11]. Moreover, in the present study, the number of lactate-fermenting bacteria was in the same range as that of lactate-oxidizing bacteria reducing humic acids in all sediment layers.
Therefore, even considering inaccuracies of MPN studies and keeping in mind that MPN counts regard only a fraction of unknown size of the total community, our results suggest that the community of humic-acid-reducing microorganisms in profundal sediments of Lake Constance consists largely of microorganisms that are unable to reduce ferric iron hydroxide directly – at least in the provided poorly soluble form – and at least some of them might represent fermenting bacteria. Unfortunately, our results do not allow us to differentiate between the lactate-oxidizing and acetate-oxidizing populations, since their respective numbers did not differ significantly in most sediment layers.
It has been shown recently that humic acids can also be reduced by sulfate-reducing, halorespiring, methanogenic, and hyperthermophilic microorganisms [10,11]. Apparently, the ability to reduce humic acids is widespread, and the spectrum of microorganisms capable of reducing iron(III) indirectly (see Section 4.2) is much larger than that of ‘iron-reducing bacteria’ sensu stricto. We conclude that humic acid reduction is a metabolic capability more common among bacteria than previously thought – a fact that should be considered in future models of electron flow in anoxic habitats.
4.2Electron shuttling via humic acids to iron(III) (hydr)oxides
The depth profiles show a decrease in the concentration of humic acids and an increase of iron concentration with increasing depth, both probably due to degradation of organic matter, calcite dissolution, and sediment compression (Fig. 1, see also ). The resulting higher density and lower porewater content of the sediment lead to an increase of the relative iron content with depth. The electron uptake (oxidation) capacities of the humic acids were 30–80 times lower than those of Fe(III) in all sediment layers, and consequently, their importance as terminal redox acceptors appears small when compared to that of the Fe(III) fraction. However, due to the ability of humic acids to shuttle electrons to Fe(III) hydroxides [2,8,12,14,27], microbial reduction of humic acids is chemically coupled to the much larger pool of oxidized iron and manganese.
Owing to their low solubility, iron and manganese oxides accumulate in sediments [28,29], where they represent one of the dominating electron acceptors, albeit endowed with a strong kinetic limitation. In order to reduce insoluble Fe(III) (hydr)oxides in the absence of electron shuttles or chelating agents, Fe(III)-reducing bacteria need to be in direct contact with the minerals, which is considered to be a rate-limiting step in iron(III) reduction. Insoluble forms of Fe(III) are reduced more slowly than soluble forms of Fe(III) . In the presence of humic acids as redox mediators, the coupled system could overcome the limitations imposed by the low electron uptake capacity of humic acids and the low solubility of the Fe(III) (hydr)oxides. As a consequence, the low electron uptake capacity does not limit the oxidation of organic substrates, and the kinetics of Fe(III) (hydr)oxide reduction can be enhanced.
It has to be mentioned, however, that humic acids extracted with NaOH at alkaline pH contain organic molecules that are already dissolved under in situ conditions but also molecules that are not dissolved at circumneutral pH. The relative contribution of either fraction to the total amount of extracted humic acids is not known, and this is true also for the potential kinetic limitations on rates of enzymatic electron transfer to solid-phase humic acids and on rates of Fe(III) oxide reduction by reduced solid-phase humic acids.
In addition to their function as electron shuttles, humic acids could stimulate microbial iron(III) reduction also by complexation of either Fe(III) or Fe(II) . The first process would render Fe(III) more accessible to the microorganisms, whereas the second would prevent adsorption of Fe(II) to the surfaces of minerals and microbial cells, thus providing sufficient ‘free surface’ for reduction . In addition, Fe(II) complexation would lower the concentration of free Fe(II), thus increasing the thermodynamic driving force for Fe(III) reduction that decreases with increasing Fe(II) concentrations. However, the actual extent of these complexing mechanisms caused by humic acids during iron reduction in natural environments still has to be determined.
The actual rate and extent of microbial reduction of humic acids and iron(III) in each layer will depend not only on the presence of the appropriate microorganisms. Also the substrate concentration and the redox potential, which is determined by the presence and electron uptake capacities of the respective electron acceptors, viz., iron(III) or oxidized humic acids, control the extent of the reduction. The specific redox capacity of the humic acids recovered from profundal sediments of Lake Constance (0.5–0.7 μequiv (mg humic acids)−1) as measured in our study corresponds to approx. 3–4% redox-active quinoid constituents in the humic acids, a value which is only slightly lower than values determined by Kappler and Haderlein  and Struyk and Sposito  for soil and aquatic humic acids, respectively.
In order to be a good redox mediator, an electron-shuttling compound must have a redox potential low enough to reduce Fe(III) and high enough to be re-reduced by the microorganisms after its chemical oxidation by Fe(III). If electrons come from the oxidation of glucose to CO2 (E°′=−434 mV) or lactate to CO2 (E°′=−343 mV) and are transferred to iron minerals with a potential between −300 and +100 mV depending on the mineral and Fe2+ concentration , a potential electron shuttle needs to have a relatively confined redox potential to fulfill this role. However, humic acids with a broad spectrum of different redox potentials indicated by nearly linear redox titration curves over a range from Eh=−300 mV to +400 mV (obtained in our lab with hexacyanoferrate as oxidant; not shown) could accomplish this demand.
Given the high apparent redox potential in the top layers (A and B) of the sediment and the large specific electron uptake capacities of the humic acids, microbial humic acid reduction is probably most favorable in these layers. However, the prevailing redox potential in the porewater (−150 to −200 mV, see Fig. 1) and the electron-accepting capacities of humic acids (0.12–0.23 μequiv (ml sediment)−1, see Fig. 3) in deeper sediment layers are probably still in the appropriate range for humic acids to serve as electron shuttles between microorganisms and iron(III) hydroxides. Both the prevailing redox potential and the electron-accepting capacity of humic acids determined in this study are comparable to the conditions given in experiments in which AQDS was added to aquifer material (E0, AQDS=−184 mV; cAQDS≤0.250μmol (g sediment)−1): in these studies, AQDS was shown to be an excellent electron shuttle between bacteria and iron(III) hydroxides . It should be emphasized, however, that the redox potentials of porewater, the humic acid pool, and the iron pool are not in equilibrium, but rather in a steady-state flow situation governed by the kinetics of specific rate-limiting reactions.
Due to the rapid chemical oxidation of Fe(II) by oxygen and the low solubility of Fe(III) in water, iron accumulates in upper sediment layers as Fe(III) (hydr)oxides . As discussed above, Fe(III) is reduced to Fe(II) below the oxic sediment surface. However, chemical reoxidation (e.g., reoxidation of iron(II) by oxygen diffusing into the sediment) and iron(II) oxidation by aerobic [35,36] or nitrate-reducing bacteria  recycle the iron(II) and thus replenish the iron(III) pool. Even in the deeper sediment layers (C and D), a complete reduction of iron(III) was not observed although reduced humic acids and a significant humic-acid-reducing bacterial population were present. It appears likely that part of the iron(III) is physically not accessible (e.g., because of formation of aggregates of iron minerals where the iron(III) minerals are covered by iron(II) precipitates) or that thermodynamic constraints imposed by the low redox potential and lower electron-accepting capacity of humic acids and/or iron(III) are responsible for the observed incomplete reduction.