The manuscript titled “Conversion of Escherichia coli to Generate All Biomass Carbon from CO2” by Gleizer et al. describes metabolic engineering and evolved Escherichia coli to produce all its biomass carbon from CO2 as a sole carbon source. This work is based on previously published two papers (Antonovsky, et al. 2016 Cell 166, 115–125) which established the Calvin cycle pathway for the synthesis of sugar from CO2 in Escherichia coil, and ; (Volpers, et al. 2016 PLOS ONE 11, 1–25) which stoichiometrically analyzed the metabolic network in E. coli and found out that the addition of FDH, Rubisco, and phosphoribulokinase (Prk) to the metabolic network of E. coli is sufficient for in silico autotrophic growth.

In this study, they established a synthetic autotrophy in E. coli using an adaptive evolution strategy for carbon fixation, and the energy is harvested from formate. The aim of this work is clear.

This article is the first work that demonstrated the fixation and utilization of CO2 in E. coli to grow autotrophically using the Calvin cycle after a very long time period of adaptation, but the main idea and novelty are based on the previous articles published in Cell (Antonovsky, et al. (2016) and in PLOS ONE (Volpers, et al. 2016).

The reviewer asks the authors to address some major comments, listed below, as well as a few minor corrections.

Major comments: 1. this experiment took a very long time for the adaptive evolution of E. coli to be able to use CO2 as a sole carbon source. Why didn't you inhibit some gene for metabolic fluxing towards the Calvin cycle which mainly mutated after the adaptive evolution? 2. This growth rate of E. coli strain growing on formate via an engineered Calvin cycle is very low. The autotrophic E. coli doubling time is 18 hours compared to 30 min for the wild type which need to be shortened. You need to test different media, add supplements or use different growth conditions. 3. RuBisCo type II used for this modified strain has a high kcat with low selectivity for CO2 over O2. In bioreactors, a high CO2 concentration can be achieved. However, for applications under atmospheric conditions or in cases in which CO2 inhibits the reaction, RuBisCO type I with a higher selectivity for CO2 was better to be used.

Minor points: 1. Figure 2 needs to be simplified. 2. The functions of other genes (category 3) that have been got from the genome sequence of adapted strain need more investigations.