Dear Zhang and co-workers,

I find this work and topic interesting as I am myself in the same field. Some comments;

1) The old work of Woodward et al was stimulating but I guessed it was unlikely to be practical given that their data showed very slow H2-production rate and transient yield-optimum. From the perspective of micro-scale energy-production (or any other scale) it is interesting to compare this approach with the recent papers of Yoshida et al in collaboration with Yukawa at RITE (http://aem.asm.org/cgi/co... and later publications), where the applied aim may be similar? They achieved very high rates of production using whole cells (approximately 10,000 times higher than what is shown in Fig. 2, 11.8 moles H2/L/hr compared to ~1 mmol H2/L/hr, unless I've mixed the numbers up..?), although the yield of formate-dependent H2-production obviously will be limited to 2 at a maximum. A discussion of practical limits with respect to production rates would have been interesting in your paper, particularly with respect to micro-scale application such as mobile phones, etc..? Yoshida et al discussed their rates in relation to a system needed to drive 1 kW PEMFC.

2) The materials and methods says helium was used to sparge the system, although Fig. 5 shows N2 as a carrier gas?? Regardless of what gas is used, I think the question of removal of H2 from any H2-producing reaction obviously will be of applied importance, as it will influence the overall reaction rate. This may be one of the obvious advantages of the cell-free system, since metabolic co-factor couples do not need to be tightly constrained by limitations set by other metabolic reactions to a particular "fixed" steady-state levels. The only research papers that I have found so far that considers removal of H2 from whole-cell production systems seems to be Dutch researchers such as Stams, Claassen, and Bussmann. In a papaer by Van Groenestijn et al (Int. J. Hydrogen Energy 27, 1141-), they estimated the energetic and economic costs of H2-removal on a large production scheme, and it appeared to have a big influence on overall cost of production. A discussion of these issues and aspects would have been interesting.

3) Following on from the above, and from your energy diagram, the reaction you studied obviously is an uphill reaction, energetically, in the absence of the connected fuel cell. Where then does this energy come from in your reaction, given that the reaction appears to proceed very well over a long time period? Is this not H2-removal? It would be nice to see a calculation of estimates of net H2-yield, after subtracting relative loss in energy (expressed as H2) as a result of energetic costs related to H2-removal (if this indeed is the major driving force?). Bruce Logan and co-workers did a similar analysis of the BEAMER-process, a process where electricity had to be applied in order to allow oxidation of acetate to take place, and then back-calculated the relative loss in potential H2 as a result of energy-loss (from the need to supply a constant current). See Liu et al. (Environ. Sci. Technol. 2005, 39, 4317-4320). What happens if you turn off the sparging and carry out the reaction under closed batch conditions?

4) It would also be interesting to see a discussion regarding the relative merit of carrying out synthetic biology reactions in vitro as opposed to in vivo. You mentioned the concept of introducing your system to a minimum genome organism, although it would be difficult to see how the outcome would be much different from introduction to a standard biotechnologically applicable microorganisms? For example, even a minimum genome organism would still most likely rely on glycolysis and glycolytic and PPP intermediates for their basic metabolic needs, and the NADP(H) co-factor couple would possibly be restrained to limit NADPH-dependent H2-synthesis to a much lower maximum partial headspace H2-pressure than what you have in the in vitro system, thereby perhaps requiring even more dilution of H2 in order to maintain sufficiently low pH2-levels to allow the reaction to proceed at reasonable rates? (If this is an issue or not is another question). Then, the possibility of ever achieving a reaction which could be regarded as near-futile (except for the loss of CO2), ie. cyclic PPP, in vivo, would be interesting to also consider. With respect to in vivo vs. in vitro, a more extensive discussion of issues and potentials for either case would be nice to see. For example, broadening your discussion of cost of enzyme production to stability and regeneration, including (in)stability of NADPH, and relating this to issues associated with in vivo H2-production?

Please excuse the long comments, but the topic is interesting and I would be interested to read your response, and yet I realize that no one paper can contain everything. Perhaps next ones?

Cheers,

Patrik