Methane Turnover in Desert Soils

Deserts cover about a third of the land surface on Earth. However, despite their size, their ecology – and particularly their microbial ecology – is far less understood than the ecology of more humid regions. Previous studies have indicated that desert soils might be involved in the production and consumption of methane, an important greenhouse gas. The turnover of atmospheric gases involves many microorganisms, and methane is no exception – it is both produced and consumed by microbes. Despite the extensive research methane has been subjected to, a rigorous study striving to elucidate methane turnover patterns in arid regions and aiming to detect the active organisms involved has not been conducted so far. This work comprises three parts. The first part deals with biogeographical patterns of soil microbial communities along a steep rainfall gradient in Israel ranging from less than 100 to more than 900 mm yr-1. We show that community profiles of both Archaea and Bacteria do not change continuously along the gradient, but rather cluster into three groups that we have defined as arid, semi-arid and Mediterranean. These three categories demonstrate a qualitative difference in the microbiology of arid soil compared to more humid regions. In the second part we show that pristine arid soils in the Negev Desert, Israel, are sinks for atmospheric methane, but that disturbed sites and pristine hyper-arid sites are probably not. The methanotrophic activity was located in a narrow layer in the soil down to about 20 cm depth. Interestingly, the biological soil crust (BSC) which is typically the most active layer in desert soils showed no methane uptake activity and was apparently devoid of methanotrophs. Transcripts of the key methanotrophic gene – encoding for the particulate methane monooxygenase (PMMO) – were detected in the active soils and their sequences showed that they are affiliated with two clusters of uncultured methanotrophs: USC and JR3. Based on a correlation of the relative abundance of each methanotroph to the methane oxidation rate we concluded that JR3 is the dominant atmospheric methane oxidizer in this arid system. The third part deals with methanogenesis in upland soils with a focus on drylands. Following previous work we show that many upland soils, sampled globally, possess a methanogenic potential, when incubated anoxically, despite being aerated most of the time. Only two active methanogens were detected – Methanosarcina and Methanocella – which appear to be universal upland soil methanogens. Under these conditions, acetoclastic methanogenesis, mediated by Methanosarcina, was the dominant methanogenic pathway and cell numbers of Methanosarcina were well correlated with methane production rates. Lastly, we show that the BSC was the source for methanogenic activity in arid soils while the deeper layers showed little or no methanogenic potential. When the BSC was incubated in a wet state in microcosms and in the presence of oxygen methanogens could still grow and methane was still produced albeit at relatively low amounts. Both methanogens expressed the gene encoding for the oxygen detoxifying enzyme catalase giving at least some explanation to their ability to remain viable in the presence of oxygen. Under these conditions, Methanocella was the dominant methanogen and most methane was produced from H2/CO2, indicating niche differentiation between the two methanogens. The findings of this work suggest that under standard dry conditions pristine arid soils are a net sink for atmospheric methane but that following a rain event they might turn into net sources