In theory, a few naturally occurring evolutionary changes in the genome of a model organism may have little or no observable impact on its wild type phenotype, and yet still substantially impact the phenotypes of mutant strains through epistasis. To see if this is happening in a model organism, we obtained nine different laboratories’ wild type Myxococcus xanthus DK1622 “sublines” and sequenced each to determine if they had evolved after their physical separation. Under a common garden experiment, each subline satisfied the phenotypic prerequisites for wild type, but many differed to a significant degree in each of the four quantitative phenotypic traits we measured, with some sublines differing by several-fold. Genome resequencing identified 29 variants between the nine sublines, and eight had at least one unique variant within an Open Reading Frame (ORF). By disrupting the ORF MXAN7041 in two different sublines, we demonstrated substantial epistasis from these naturally occurring variants. The impact of such inter-laboratory wild type evolution is important to any genotype-to-phenotype study; an organism’s phenotype may be sensitive to small changes in genetic background, so that results from phenotypic screens and other related experiments might not agree with prior published results or the results from other laboratories.

Information about a gene sometimes can be deduced by examining the impact of its mutation on phenotype. However, the genome-scale utility of the method is limited because, for nearly all model organisms, the majority of mutations result in little or no observable phenotypic impact. The cause of this is often attributed to robustness or redundancy within the genome, but that is only one plausible hypothesis. We examined a standard set of phenotypic traits, and applied statistical methods commonly used in the study of natural variants to an engineered mutant strain collection representing disruptions in 180 of the 192 ABC transporters within the bacterium Myxococcus xanthus. These strains display continuous variation in their phenotypic distributions, with a small number of “outlier” strains at both phenotypic extremes, and the majority within a confidence interval about the mean that always includes wild type. Correlation analysis reveals substantial pleiotropy, indicating that the traits do not represent independent variables. The traits measured in this study co-cluster with expression profiles, thereby demonstrating that these changes in phenotype correspond to changes at the molecular level, and therefore can be indirectly connected to changes in the genome. However, the continuous distributions, the pleiotropy, and the placement of wild type always within the confidence interval all indicate that this standard set of M. xanthus phenotypic assays is measuring a narrow range of partially overlapping traits that do not directly reflect fitness. This is likely a significant cause of the observed small phenotypic impact from mutation, and is unrelated to robustness and redundancy.

The Clp/Hsp100 proteins are chaperones that play a role in protein degradation and reactivation. In bacteria, they exhibit a high degree of pleiotropy, affecting both individual and multicellular phenotypes. In this article, we present the first characterization of a Clp/Hsp100 homolog in Myxococcus xanthus (MXAN_4832 gene locus). Deletion of MXAN_4832 causes defects in both swarming and aggregation related to cell motility and the production of fibrils, which are an important component of the extracellular matrix of a swarm. The deletion also affects the formation of myxospores during development, causing them to become sensitive to heat. The protein product of MXAN_4832 can act as a chaperone in vitro, providing biochemical evidence in support of our hypothesis that MXAN_4832 is a functional Clp/Hsp100 homolog. There are a total of 12 Clp/Hsp100 homologs in M. xanthus, including MXAN_4832, and, based on its mutational and biochemical characterization, they may well represent an important group. The Clp/Hsp100 proteins are ubiquitous among bacteria, lower eukaryotes, and plants. At the molecular level, Clp/Hsp100 proteins are all thought to be chaperones that function as protein reactivators and/or degradation partners (8,12). At the cellular level, the function of Clp/Hsp100 proteins is complicated, as evidenced by the fact that these proteins exhibit a high degree of pleiotropy (25). For example, in Bacillus subtilis, the Clp/Hsp100 family member ClpC was found to regulate genetic competence (17), whereas, in the pathogen Listeria monocytogenes, a ClpC paralog is involved in escaping from phagosomes (23, 24) and in host cell adhesion and invasion (21). This pleiotropy has been observed not only between species but also between Clp/Hsp100 homologs within the same species; a second Clp/Hsp100 homolog in L. monocytogenes, ClpE, is involved in both virulence and cell division (20). From the accumulated data, it seems logical to postulate that members of the Clp/Hsp100 family of proteins function both in a bacterial cell's individual cellular processes and in its interactions with the environment.No Clp/Hsp100 protein has been characterized in the multicellular prokaryote Myxococcus xanthus, even though there are ample opportunities for these proteins to play a role in its complex response to starvation stress. On nutrient-rich agar, M. xanthus grows as a single-species biofilm called a swarm. On nonnutritive (starvation) agar, a swarm does not grow, and the organism instead executes a stepwise developmental process that culminates in the formation of multicellular dome-shaped aggregates called fruiting bodies. When development is completed, the swarm's millions of cells are parsed into fruiting bodies of ϳ1 ϫ 10 5 cells, a subset of which have differentiated to become metabolically quiescent and environmentally resistant myxospores located in the center. Development requires control of cell movement, division, and differentiation, as well as coordination and communication between cells and with the environment. C...