INTRODUCTION
Root-nodule symbioses (RNS) between plants and nitrogen (N) fixing bacteria are ecologically and economically important interactions (Graham & Vance, 2003; Vitousek et al. 2002; Wheeler & Miller, 1990), and provide model systems for studying the evolutionary ecology of mutualism (Heath & Grillo, 2016). RNS are phylogenetically widespread, occurring across a diverse set of organisms that include the legume plants (Fabaceae) and Parasponia spp. that form ‘rhizobial’ associations with bacterial symbionts from three proteobacterial phyla, a different set of plants from 8 families and 25 genera that form ‘actinorhizal’ associations with the single actinomycete genus Frankia , and some species of cycads that associate with cyanobacteria (Dawson, 2008; Piex et al . 2015; Vessey et al . 2005). In all RNS, host plants acquire bacterial symbionts ‘horizontally’; i.e., plants form nodules de novo with bacteria found in the soil environment, rather than receiving symbionts ‘vertically’ from their parents, as in some endosymbiotic interactions (Denison & Kiers 2011; Frank, 1996; Wall & Berry, 2008). The independent existence of plant and bacterial partners in RNS affects the evolution of the symbiotic interaction in several ways.
One result of this independence is the potential for evolutionary conflict between host plants and symbiotic bacteria. Because individual plants associate simultaneously with bacteria from multiple lineages that can vary widely in the benefits they provide to the host (Denison, 2000; Kiers and Denison, 2008; Markham, 2008; Parker, 1995; Sachset al ., 2010), a Tragedy-of-the-Commons can occur in bacterial communities (Denison 2000), favoring potentially ‘selfish’ behavior such as decreased allocation of host-derived carbon to N-fixation (Oono & Denison, 2010), or bacterial reproduction within host nodules that evades host regulation (Cotin-Galvan et al ., 2016). Over evolutionary time, selection for such ‘cheating’ behavior should destabilize the mutualism, in the absence of other factors (Sachs and Simms, 2006). However, bacterial cheating is thought to be held in check by several countermeasures evolved by host plants. A multi-level system of pre-nodulation communication mechanisms (Clúa et al ., 2018) between roots and bacteria allows the plant to recognize bacterial genotypes likely to provide effective N-fixation (‘partner choice’ Kiers & Denison, 2008). Post-nodulation detection of bacterial N fixation allows plants to ‘punish’ cheating genotypes via decreased reproduction within nodules (termed ‘host sanctions’ (Denison 2000)) and/or selectively allocate resources to especially beneficial bacterial genotypes (‘differential rewards’ Simms & Taylor, 2002).
Horizontal transmission of RNS symbionts also means that the bacterial partner occupies two distinct niches: one in which they live actively or dormantly in soil, often forgoing N-fixation entirely, and one in which they essentially act as plant organelles, fixing N using host-derived carbon sources that, in the most effective symbiotic genotypes, are allocated preferentially to N-fixation over bacterial reproduction (Benson & Silvester, 1993; Denison & Kiers 2004). These two niches are likely to exert different selection pressures on N-fixing bacteria, which may result in tradeoffs between symbiotic and free-living lifestyles (West et al ., 2002) and/or drive repeated gains and losses of symbiotic capability in bacterial lineages (Sachs et al . 2010; Sachs & Simms, 2006). Among the symbiotic fraction of the bacterial community, passage between soil-dwelling and nodule-dwelling phases has the potential to exert strong but complex influences on the mixture of genotypes available to host plants. In order to be available for nodule-formation in the first place, bacterial genotypes must be able to maintain a viable existence in soil. Once chosen by a host plant, a bacterial cell can be cloned to much greater densities within a nodule before re-entering the soil during nodule senescence (Denison 2000), increasing the proportional representation of the chosen genotype in the soil fraction, as well as its availability for future rounds of nodule formation.
Little is known about the relationship between nodule and soil assemblages of N-fixing bacteria in natural environments. Historically, investigation of diversity patterns in soil-dwelling bacterial symbionts have been based on the use of N-fixing ‘trap plants’ to sample bacterial communities via nodule formation or, less frequently, on direct isolation of bacteria from soils (Chaia et al ., 2010; McInneset al ., 2004). Both approaches are subject to significant and well-known biases (Chaia et al ., 2010; Kirk et al . 2004). In recent years, molecular methods such as DNA cloning, qPCR, and next-generation sequencing (NGS) methods have been increasingly used to characterize soil communities of N-fixing bacteria (e.g., Ben Tekayaet al . 2017, 2018; Miranda-Sanchez et al . 2016, Rodriguezet al ., 2016). While these methods are not without their own biases (Acinas et al ., 2005; Chandler et al ., 1997; Gołębiewski & Tretyn, 2020; Sáenz et al ., 2019; Sanchez-Cidet al ., 2022; van Elsas & Boersma, 2011), they allow investigation of a much broader swath of the bacterial community not limited to genotypes able to induce host nodules and/or grow under laboratory conditions. In the present study, we use PCR and DNA cloning to examine patterns of genetic diversity in naturally-occurring soil-dwelling populations of the actinorhizal symbiont Frankiaspp. associated with the host species Alnus incana ssp.tenuifolia (hereafter ‘A. tenuifolia ’) across a primary successional sere in the boreal forest of Alaska. We utilize a subset of research sites in which we have previously observed stable nodule-dwelling assemblages over a range of 3 to 5 years, with large differences in composition among successional stages that are consistent among replicate sites (Figure S1) (Anderson et al ., 2009, 2013; Ruess et al . 2013). In the present study, we investigated three questions: 1) how do soil and nodule assemblages compare in phylogenetic diversity of Frankia , 2) do soil and nodule assemblages change in parallel in response to environmental variation: both across succession, and in response to long-term fertilization with nitrogen (N) and phosphorus (P), and 3) do soil assemblages differ in the presence and absence of A. tenuifolia individuals?