Multivariate analysis of soils: microbial biomass, metabolic activity and bacterial community structure and their relationships with soil depth and type

A multivariate statistical approach based on a large data set of abiotic and biotic variables wasused to classify four contrasting-land-use soils. Soil samples were collected at increasing depthfrom a calcareous agricultural soil, a temperate upland grassland soil, a moderately acidic agriculturalsoil, and an acidic pine forest soil. Analytical investigations were carried out by using acombination of conventional physical, chemical, and biochemical methods coupled with denaturinggradient gel electrophoresis (DGGE) community fingerprinting of PCR-amplified 16S rRNAgene-coding fragments from soil-extracted total-community DNA. The data set of soil physical,chemical, and biochemical variables was reduced in dimensionality by means of a principalcomponent-analysis (PCA) procedure. Compositional shifts in soil bacterial-community structurewere analyzed through a clustering algorithm that allowed identifying six main bacterial-communityclusters. DGGE fingerprinting clusters were further analyzed by discriminant analysis (DA)using extracted PCA components as explanatory variables. Soil organic matter–related pools(TOC, TN) and functionally related active pools (microbial biomass C and N, K2SO4-extractableC) significantly decreased with soil depth, and resulted statistically linked to one other and positivelyrelated to enzymatic activities (acid phosphatase, arylsulfatase, b-glucosidase, dehydrogenase,hydrolysis of fluorescein diacetate) and silt content. Besides organic-C gradients, pedogenetic-driven physico-chemical properties, and possibly soil thermal and moisture regimesseemed to play a key role in regulating size and energetic ecophysiological status of soil microbialcommunities. DGGE analysis showed that contrasting horizons were conducive to the dominanceof particular bacterial ribotypes. DA revealed that the bacterial-community structure wasmainly influenced by organic matter–related variables (TOC, TN, CEC, Cflush, Nflush, Extr-C),chemical properties such as pH, CaCO3, and EC, together with textural properties. Results indicatethat, beyond land use or plant cover, pedogenetic-driven physico-chemical conditions changingwith soil type and depth are the key factors regulating microbial size and activity, and determiningthe genetic structure of bacterial community.