As for the rate of synthesis, one indirect way to consider the problem that I eluded to, may be the stringent response. When even a single amino acid is missing from media, uncharged tRNA are formed and trigger ppGpp synthesis by RelA. This induces a stress response resulting in large downregulation of the whole protein transcriptional/translational machineries that apply strong limits on growth, as in the stationary phase. If amino acid synthesis (and its kinetics) was enough in minimal medium (regardless of whether the energy costs of biosynthesis are positive or negative), this response would likely not occur.

These references may be helpful:

http://www.blackwell-synergy.com/doi/abs/10.1111/j.1365-2958.2008.06229.x

http://jb.asm.org/content/181/20/6425.short

Unfortunately, and as Kaspar mentioned, I can’t recall a study that clearly compares the costs of amino import vs biosynthesis, although there might be more hints in the literature. While there must be costs in producing transporters and importing amino acids from rich media, these costs should be much lower than biosynthesis which requires multiple enzymes for each amino acids and large scale reorganisation of activities.

This modelling analysis is also interesting:

http://www.pnas.org/content/99/6/3695.short

This study and others have examined how the biosynthetic costs of amino acid synthesis seemed to have biased the amino acid composition of abundant proteins in E. coli and other organisms.

Dear Karl, I think amino acids provide cell not only with required C, but also with highly valuable N required for protein and nucleic acid biosyntheses, which are required for growth and development. Amino acids are among small soluble organic molecules that can easily absorbed from the surrounding environment.

Thank you for the answers, I am sorry if my question was not very precise. I meant to ask why are cells able to grow faster with amino acids compared to glucose minimal medium (supplied with exess amounts of N, S and all other micro elements, I think E. coli is not able to grow solely on glucose.)

Also if the N and S is used for protein biosynthesis why not use the amino acids directly for protein synthesis (instead of metabolising them first into N and S compounds).

I think if enhanced growth rates are achieved as / coupled to increased protein synthesis and if increased aminoacid abundance signals more protein biosynthesis, then the presence of aminoacids supplements will speed up the growth.

Dear Karl for E. coli medium as You know yeast extract is applied as an organic source of various amino acids. I think this is a better way as there may be various auxotrophic races of E. coli that are not able to synthesize certain amino acids. With the rapid generation of bacteria the appearance such races is highly expected. Then if the aim is just to grow the bacteria, yeast extract and peptone will properly do the task as cheaper complex sources of amino acids. Additionally, pure amino acids are more expensive.

I am talking about lab experiments with single strain non auxotroph E. coli that can either synthesise amino acids or transport them. The question is when you grow cells on minimal medium vs minimal medium + amino acids. What is the growth advantage? This paper by Kaleta shows that when cells synthesise amino acids they gain ATP. Therefore when they import necessary amino acids for protein synthesis they should produce less ATP compared to synthesising these amino acids from glucose + N and S salts. Where does the growth advantage come from? Is the transport of amino acids less costly than the transport of glucose + N and S salts?

Dear Karl,

Whatever the ATP costs of amino acid biosynthesis are, the cellular investments in biosynthetic machinery (enzymes for amino acid production) and the rate of their biosynthesis are important factors limiting growth rate on minimum medium.

If I'm not mistaken, the Kaleta et al, paper does not consider these factors. In addition, the absence of amino acids induces the stringent response which triggers large gene expression profile changes that limit growth.

Thank you Martin,

I agree with you on the biosynthetic machinery investment, this was not considered by Kaleta et al (and is quite difficult to do). Although this means that the investment into amino acid transport enzymes has to be lower than to biosynthesis which could easily be true.

I hadn't thought on the limiting rate of amino acid biosynthesis.

Hello Karl,

I think that a possible cause is : if amino acids are present in the medium, the bacterium ceases to synthetises it and directly uptake the AA. This las process constitutes a non negligible energy gain for the cell. This energy gain can be invested in the cell biomass production ... and the growth rate increases.

Hi Jacques,

Your answer contradicts with the results of Kaleta et al 2013 I referred to in the question. They show that the biosynthesis of amino acids results on average in production of ATP. Therefore no biosythesis of amino acids results in loss of ATP compared to the first situation.

I understand that there is also the investment the protein machinery for AA biosynthesis. This then means that the investment for AA biosynthesis proteins has to be lower compared to the proteins necessary for import of amino acids.

Hi,

I concur with Martin. It should be the protein costs that cells save which makes the difference. I also expect that the cost of a transporter (in total amounts) is much smaller than the protein cost of the corresponding amino acid biosynthetic pathway (at least when the amino acid is present in sufficient quantities).

Hello Karl,

You wrote: " This then means that the investment for AA biosynthesis proteins has to be lower compared to the proteins necessary for import of amino acids."

Why not ? Is the contrary demonstrate ?

I don't know of such a data set yet confirming or over-ruling this assumption since absolute quantification of all amino acid transport and biosynthesis proteins in one study is very difficult mostly due to the fact that transport proteins are membrane-bound.

But I would agree with Karl, Martin and Christoph that the protein synthesis cost of one (or a couple) transport protein(s) is much smaller that the one of the whole biosynthesis pathway. And this would also be a possible answer to Karl's question together with the possible rate limitation of amino acid biosynthesis.

Jacques, I might have misunderstood your sentence " amino acids are present in the medium, the bacterium ceases to synthetises it and directly uptake the AA" if by this you mean the energy gain comes from the proteome then I agree. It is important to differentiate between the ribosome load/energy for proteome synthesis and energy directly for amino acid synthesis as already stated in previous posts...

Hi Karl,

I know that my English is not so good so it is normal to misunderstake some of my sentences. Finally, many thanks for the rectification: its an honor !

Martin, could you share some literature showing that the rate of amino acid biosynthesis could be the growth limiting factor on minimal medium? Or at least somehow even indirectly showing that removal of this limit enables cells to grow faster when they can obtain amino acids from the environment?

Cheers!

As for the rate of synthesis, one indirect way to consider the problem that I eluded to, may be the stringent response. When even a single amino acid is missing from media, uncharged tRNA are formed and trigger ppGpp synthesis by RelA. This induces a stress response resulting in large downregulation of the whole protein transcriptional/translational machineries that apply strong limits on growth, as in the stationary phase. If amino acid synthesis (and its kinetics) was enough in minimal medium (regardless of whether the energy costs of biosynthesis are positive or negative), this response would likely not occur.

These references may be helpful:

http://www.blackwell-synergy.com/doi/abs/10.1111/j.1365-2958.2008.06229.x

http://jb.asm.org/content/181/20/6425.short

Unfortunately, and as Kaspar mentioned, I can’t recall a study that clearly compares the costs of amino import vs biosynthesis, although there might be more hints in the literature. While there must be costs in producing transporters and importing amino acids from rich media, these costs should be much lower than biosynthesis which requires multiple enzymes for each amino acids and large scale reorganisation of activities.

This modelling analysis is also interesting:

http://www.pnas.org/content/99/6/3695.short

This study and others have examined how the biosynthetic costs of amino acid synthesis seemed to have biased the amino acid composition of abundant proteins in E. coli and other organisms.

Another factor to put into this discussion is rate instead of just energy. While I'm not sure that I agree with the cited article and all of its underlying assumptions, even were you to accept it then there is still a rate issue. Actively growing cells with a full contingent of ribosomes translate protein at a rate of 100,000-400,000 residues per sec per cell. That means 100K-400K molecules of amino acid are needed per cell per sec. I would guess (but I haven't done the math) that this would consume an inordinate amount of available precursors and energy such that the mere rate of synthesizing amino acids is limiting for maximal cell growth. So the reduced cell growth seen in minimal medium without amino acids might not be an energy issue but rather just a reflection of how fast the cells can make high levels of amino acids.

To Martin:

Thanks for the references!

Do I understand your stringent response idea correctly that by "This induces a stress response resulting in large downregulation of the whole protein transcriptional/translational machineries..." you mean that in minimal medium, the fraction of total protein synthesis capacity for amino acid synthesis is higher than the fraction of amino acid transport and metabolism in amino acid supplemented medium? Which would mean that basically less translational resource is available for synthesizing everything else in minimal medium e.g. ribosomes, central metabolism enzymes etc.? Still, this would not explicitly solve the rate issue... More so, if we know that the catalytic rates of enzymes can vary strongly under different conditions.

Indeed, like Akashi and Gojobori 2002, we saw in our recent paper with absolute quantitative proteomics analysis of E. coli (Valgepea et al. 2013) that more abundant proteins cost less costly amino acids, and vice versa. However, I used the amino acid biosynthesis costs from Akashi and Gojobori, and since the costs in Kaleta et al. 2013 are very different from the latter it would be interesting to repeat the analysis.

To Michael:

Still on the rate issue. I totally understand it intuitively that synthesizing an amount of amino acids takes longer than to transport them but like I already ask Martin, I would love to see some experimental evidence for this - showing that the rate of amino acid biosynthesis could be the growth limiting factor on minimal medium?

@Kaspar

Concerning your first point and this of Martin I do concur. In minimal medium a considerable part of the proteome is invested into enzymes for amino acid biosynthesis while in rich medium they are not needed and other proteins (e.g. for translation) can be produced, thereby improving growth rate.

Concerning your second point, we actually did a similar analysis like in the Akashi and Gojobori paper with our costs and a more recent protein abundance data set (approx. 750 proteins on minimal medium). We saw similar trends though somewhat weak. We didn't include it in the manuscript due to space constraints.

Concerning your third point I'm not sure whether amino acid biosynthesis (i.e. the conversion of carbon source into amino acids) itself is really rate limiting. There is a paper by Stoebel (http://www.ncbi.nlm.nih.gov/pubmed/18245823) in which it is shown that amino acid supply is not really rate limiting during overexpression of the lac operon.

To Christoph:

1) Indeed, amino acid transport and metabolism proteins (COG group E) account for ~11 and ~15% from the total proteome of E. coli based on protein concentration or protein synthesis cost (concentration x length x polymerization cost), respectively (Fig. S8 in Valgepea et al. 2013). This data supports the hypothesis proposed by Martin and Christoph that supplying cells with amino acids frees up this fraction and cells could use this for increasing proteome fractions devoted to improving maximum specific growth rate e.g. translation. However, we currently preparing a manuscript based on experimental data which does not support this explanation well - the difference in protein synthesis cost of COG group E proteins is only 1% lower in rich defined (supplemented with 20 AAs) compared to minimal medium in E. coli. I found this very surprising.

2) Did you use E. coli data? Which data set? I am interested since we saw significant effects with Akashi and Gojobori costs and our proteome data set for 1,100 E. coli proteins.

3) I remember this paper now and indeed it alludes that amino acid synthesis rate may not be growth limiting after all. I am not sure though how directly can the results of lac operon overexpression be translated to the "normal" growth state. Because if neither the "proteome allocation" nor the synthesis rate hypothesis explain the faster growth in rich compared to minimal medium, then what does?

Christoph, I don't quite agree with your interpretation of that paper. The authors conclude that the cost of constitutive expression arises from the cost of expression and not from the cost of using the amino acids. That part is correct. But in their experiments the amino acids were equally limiting (or not) under all conditions. They never tested that aspect of the hypothesis, just whether the cost of expression was due to using amino acids. It isn't quite the same thing.

@Kaspar:

1) I'm not sure whether I understood this point correctly. Do you state that the production cost of the synthesized proteins on amino acid supplemented medium is only slightly lower than in a medium in which they have to synthesize the amino acids themselves? In which units are these costs calculated?

2) The data is an expanded version of some earlier protein abundance data from E. coli on glucose-minimal medium and can be found in the Supplementary Data Set of this article http://onlinelibrary.wiley.com/doi/10.1038/msb.2011.46/suppinfo (second item).

3) Has the "freed-up-proteom"-hypothesis been refuted? Also in some (unpublished) experiments with E. coli in which we provided individual amino acids to glucose minimal medium we saw a strong down-regulation of the corresponding proteins (based on 2D-gel-data).

@Michael:

Okay, I agree that they didn't test is directly. It is more of a general impression that amino acid biosynthesis in E. coli has some reserve capacity. Also we've frequently seen that amino acids are actually excreted during growth in batch medium even by wild-type E. coli (they are taken up again before entry into the stationary phase). At the end I guess it is difficult to consider any factor limiting for growth as it's more of a compromise between the allocation of proteomic ressources to the protein biosynthetic machinery, biosynthetic proteins and a lot of other stuff.

Despite the theory, the observation remains that E. coli grows significantly slower in the absence of added amino acids than in its presence. This is true whether or not the strain undergoes stringent response, or is a relA mutant.

It might have to do with the cost of uptake and metabolism, but that just doesn't seem intuitive to me. Although they comprise a significant portion of the proteome, I believe the bulk of those proteins are present regardless of the presence or absence of exogenous amino acids.

@Christoph:

1) Yes, you understood correctly - the total protein synthesis cost (concentration x length x polymerization cost; same calculation as in Valgepea et al. 2013) of COG group E proteins is only 1% lower in rich defined (supplemented with 20 AAs) compared to minimal medium in E. coli.

The latter result alludes that the "proteome allocation" hypothesis does not explain the different specific growth rates in rich and minimal medium. Actually, Stoebel et al. 2008 data supports that by stating "Sequestering amino acids in useless protein is not a major cause of selection against constitutive expression". As actually indeed the Stoebel et al. 2008 paper does not explicitly allow to reject the "amino acid synthesis rate limitation", then it still seems that amino acid synthesis rate is the more probable cause. I would love to see experimental data testing this.

2) It is based on the Lu et al. 2007 data with 450 proteins. I am very interested what the results would be with our latest 1,100 protein data set.

The say the bioenergetic costs for 'many' but not for all amino acids are negative. If you have to operate the synthesis machniery for specific amino acids as aromatic once, histidine or tryptophane the costs are highly positive. Aditionally, the nitrogen source is important. I guess at the moment there is no comprehencive model to predict the real costs for all amino acids needed for protein synthesis.

And if you look to the experiments you can measure the difference in the growth rate.

@Kaspar: I find it quite strange that protein costs for amino acid biosynthetic genes are almost the same in rich and minimal medium. In rich medium there is a strong down-regulation of amino acid biosynthetic proteins so the costs should be smaller or am I mistaken?

@Kaspar

Your interpretation of one of the outcomes of the stringent response may be correct but what I meant was simply that stringent cells shut down most growth processes. As Michael pointed out however, a relA mutant, not displaying downstream stringent response effects, but still growing slowly in absence of amino acid suggests that the limitation is elsewhere, such as tRNA charging.

As I tried to express before, to me the mere fact that uncharged tRNA appear in minimal medium and absence of amino acids suggests some imbalance and biophysical limitation in amino acid bioavailability, either due to kinetics of their biosynthesis and/or nitrogen assimilation, or the like.

The following paper Elf and Ehrenberg (2005), while theoretical is supportive of this possibility and shows hypersensitivity of aminoacylation reactions to variation in the rate of amino acid supply. They also show an interesting link between the rate of supply of an amino acid and its codon frequency.

http://www.sciencedirect.com/science/article/pii/S0006349505730939

It’s a long and more complex paper but they cite many relevant papers.

@Christoph:

Like I said, I was surprised myself. But the data clearly shows that. As I have not seen a data set of comparing the protein synthesis costs of amino acid transport and metabolism proteins (COG group E) in minimal and defined amino acid supplemented medium at identical physico-chemical conditions for E. coli, I cannot comment on your remark. I would be grateful if you could point me to such a data set.

I note that it is very important that the physico-chemical conditions are identical in such a comparison since, for example, if you would do this comparison in the exponential phase of batch growth, cells in rich medium will be growing much faster (as we are discussing here) and hence they will already from that factor be having different concentrations of COG group E proteins. In addition, I would like to remind you that I was speaking about protein synthesis cost, not their levels.

@Martin:

A paper increasing my votes for the hypothesis of amino acid synthesis rate. Thanks!