(B-67) Efficient optimization of transformation protocols for oral streptococci using response surface methodology

Introduction: Many streptococcal bacteria show potential for induced competence, whereby bacteria are forced to take up and recombine extracellular DNA. Induced competence can be used to genetically characterize and manipulate these bacteria. Existing protocols for inducing competence in streptococci have not been optimized for maximal transformation efficiency. Here, we apply response surface methodology (RSM) to efficiently screening for improved transformation protocols in several streptococci from the oral microbiome. Our approach leverages a central composite design that simultaneously varies multiple factors, resulting in models that identify individual factors and interactions that affect transformation efficiency. We show that each species of streptococci responds differently to the timing and reagent concentrations in a transformation protocol. Fortunately, statistical models allow us to tailor a general transformation protocol for each species. Understanding both the factors that impact the transformation efficiency and the differences between species will improve methods for studying and manipulating these bacteria.


Materials and Methods: S. mutans UA159, S. gordonii CH1, and S. sanguinis SK36 were transformed with pRW17 and plated with or without selection to assess transformation efficiency. We separately tested two peptide pheromones that induce competence in these bacteria: competence stimulating peptide (CSP) and a comX-inducing peptide (XIP) specific to S. mutans. CSP protocols were carried out in THY and the XIP protocol was carried out in CDM.

The general transformation protocol follows (with ranges following each factor in parentheses): 1) Dilute an overnight culture 1:40 in appropriate growth media. 2) Monitor the cultures until an optical density (OD) of 0.15 (0.1, 0.2). 3) Add 50 μg (5 μg, 500 μg) CSP or XIP per mL of culture. 4) Incubate for 20 minutes (10 minutes, 30 minutes). 5) Add 50 ng of DNA (pRW17) to 500 μL of culture. 6) Incubate for 60 minutes (30 minutes, 120 minutes). 7) Perform serial dilutions and spot onto both control and selection plates. 8) Count colonies the next day and estimate transformation efficiency.

The four factors, OD, CSP or XIP concentration, first incubation time, and second incubation time were investigated with a VKM central composite design. Liquid handling robots were used to perform serial dilutions and spot plate serial dilutions to improve efficiency of the protocol. Transformation efficiency was calculated from the number of colony forming units (CFUs) on the selection plates relative to the total CFU counts on non-selective plates. Statistical analyses, including linear modeling, Box-Cox analysis, and model diagnostics, were carried out in R.


Results, Conclusions, and Discussions: A response surface model (RSM) was created with data collected for each species/induction pheromone pair. All models showed constant variance of residuals, normally distributed error, and no indication of outliers within the datasets, indicating valid interpretations can be made with these models. Each model revealed differences in the factors that affect competence in each species. For example, increased OD increases transformation efficiency for S. mutans when induced with XIP but decreases it when induced with CSP. Delay 2 had a large significant effect in transformation efficiency across all species, suggesting that increasing the second incubation duration would boost transformation.

Our models also revealed species-dependent interactions between some factors. S. mutans has only two factors that interact significantly (CSP concentration and delay 2), but S. gordonii shows significant interactions between all factors except for the interacting terms in S. mutans. These results suggest differences in the way these two species process competence signals from CSP.

Overall, our results suggest that transformation protocols must be tuned for each species and pheromone to achieve maximum transformation efficiency. RSM offers a fast and structured way for identifying factors that affect transformation efficiency. More importantly, the quantitative models generated by RSM allows experimenters to computationally search the multivariate design space for improved protocols.