Escherichia coli - ACS Publications - American Chemical Society

Subscriber access provided by Binghamton University | Libraries

Article

Engineered assimilation of exogenous and endogenous formate in Escherichia coli Oren Yishai, Leander Goldbach, Hezi Tenenboim, Steffen N. Lindner, and Arren Bar-Even ACS Synth. Biol., Just Accepted Manuscript • Publication Date (Web): 30 May 2017 Downloaded from http://pubs.acs.org on June 1, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Synthetic Biology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Engineered

ACS Synthetic Biology

assimilation

of

exogenous

and

endogenous

formate

in

Escherichia coli Oren Yishai, Leander Goldbach, Hezi Tenenboim, Steffen N. Lindner, Arren Bar-Even* Max Planck Institute of Molecular Plant Physiology, Am Mühlenberg 1, 14476 Potsdam-Golm, Germany * corresponding author: Phone: +49 331 567-8910 Email: [email protected]

Key words: formate metabolism, one-carbon metabolism, auxotrophic strains, threonine cleavage, reductive glycine pathway, serine-threonine cycle, formate-THF ligase, pyruvate formate-lyase

ACS Paragon Plus Environment 1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 28

Abstract Decoupling biorefineries from land use and agriculture is a major challenge. As formate can be produced from various sources, e.g., electrochemical reduction of CO2, microbial formate-assimilation has the potential to become a sustainable feedstock for the bioindustry. However, organisms that naturally grow on formate are limited by either a low biomass yield or by a limited product spectrum. The engineering of a model biotechnological microbe for growth on formate via synthetic pathways represents a promising approach to tackle this challenge. Here, we achieve a critical milestone for two such synthetic formateassimilation pathways in Escherichia coli. Our engineering strategy involves the division of the pathways into metabolic modules; the activity of each module – providing at least one essential building block – is selected for in a designated auxotrophic strain. We demonstrate that formate can serve as a sole source of all cellular C1-compounds, including the beta-carbon of serine. We further show that by overexpressing the native threonine cleavage enzymes, the entire cellular glycine requirement can be provided by threonine biosynthesis and degradation. Together, we confirm the simultaneous activity of all pathway segments of the synthetic serine‒threonine cycle. We go beyond the formate bio-economy concept by showing that, under anaerobic conditions, formate produced endogenously by pyruvate formate-lyase can replace exogenous formate. The resulting prototrophic strain constitutes a substantial re-wiring of central metabolism in which C1, glycine, and serine metabolism proceed via a unique set of pathways. This strain can serve as a platform for future metabolic-engineering efforts and could further pave the way for investigating the plasticity of metabolic networks.


Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Bioprocesses account for merely 3% of the current production of commodity and specialty chemicals.1 The long-term goal of displacing fossil carbons with biorefineries as the prime source of fuels and value-added chemicals depends on identifying suitable feedstocks that can sustain production at hundreds of millions of tons per year.2 The use of simple sugars and starch directly competes with human consumption and hence can undermine food security and decrease biodiversity.3-4 Lignocellulosic and algal biomasses are less problematic for food security but are also limited in availability and are difficult to process.5-6 An ideal feedstock would therefore be a compound that can be produced from multiple highly available sources and can be easily handled. In previous studies we put forward the case for formate as such a promising feedstock,7-9 as it can be efficiently derived from electrochemical reduction of CO2,10-15 photoreduction of CO2,16 hydrogenation of CO2,17-18 selective oxidation of biomass,19-21 partial oxidation of natural gas,22 and hydration of syngas (i.e., carbon monoxide).23 However, most biotechnological organisms cannot naturally grow on formate. Those that can are either limited by the use of ATP-intensive pathways, resulting in low biomass and product yields (e.g., Cupriavidus necator uses the Calvin−Benson cycle24), or are difficult to cultivate, manipulate and engineer, resulting in a very narrow product spectrum (e.g., acetogens and methanogens25-26). To successfully realize the formate bio-economy concept, it is therefore advantageous to focus on model organisms for which a substantial amount of metabolic knowledge and a well-developed geneticengineering toolset are available. Moreover, as these organisms have already been optimized for industrial applications they can support the production of a wide array of value-added chemicals. Natural formate-assimilation pathways begin by condensing formate with tetrahydrofolate (THF) to generate 10-formyl-THF, which is then reduced to 5,10-methylene-THF.9 In previous studies we suggested several synthetic routes that can convert 5,10-methylene-THF into a central-metabolism intermediate and could thus sustain growth on formate in model microbes.7-9 Two of these promising pathways are the reductive glycine pathway and the serine−threonine cycle (Figure 1). The reductive glycine pathway is short, linear, and with limited overlap with central metabolism (Figure 1A). It further represents the most efficient aerobic route for formate assimilation (in terms of ATP consumption and hence biomass yield).7-9 Within the pathway, the reversible glycine cleavage system condenses 5,10-methylene-THF with CO2 and NH3 to produce glycine, which is further condensed with an additional 5,10-methylene-THF to generate serine. Serine is deaminated to pyruvate which can sustain biomass production. Overall, two formate molecules and one CO2 molecule are metabolized to a central C3-metabolite, consuming three NAD(P)H and two ATP molecules. The serine−threonine cycle, on the other hand, is longer, cyclic, and considerably overlaps with endogenous metabolism (Figure 1B). Within this cycle, glycine is condensed with 5,10-methylene-THF to give serine, which is deaminated to pyruvate, then carboxylated to oxaloacetate, and further metabolized to threonine. Threonine is then cleaved to regenerate the glycine, while producing acetyl-CoA as biomass precursor (e.g., glyoxylate shunt). Overall, one formate molecule and one CO2 molecule are metabolized to a central C2-metabolite, consuming 3 NAD(P)H and 5 ATP molecules.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 28

The serine−threonine cycle represents an adaptation of the natural serine pathway – operating in various methylotrophs27 – to the endogenous metabolism of model microbes such as Escherichia coli and Saccharomyces cerevisiae. Specifically, this cycle bypasses the formation of the toxic intermediate hydroxypyruvate28 and recycles glycine via threonine biosynthesis/cleavage instead of using the uncommon enzymes malyl-CoA synthase and lyase. The implementation of the serine−threonine cycle in model microbes is expected to be substantially easier than that of the native serine cycle, due to a lower interference with central-metabolic fluxes and avoidance of reactive intermediates.9 We aim to implement the reductive glycine pathway and the serine−threonine cycle in the model microbe E. coli. The genome of this bacterium naturally encodes all the enzymes of both pathways, with the exception of formate-tetrahydrofolate ligase (FTL), which was shown to be fully active when recombinantly expressed.29 However, the presence of the necessary genes in E. coli’s genome does not guarantee that their expression pattern and enzymatic properties are suitable to support efficient pathway activity. Also, overlap of the synthetic pathways with endogenous metabolism might constrain the required fluxes to support formate assimilation. Moreover, the key enzymes of the synthetic formate assimilation pathways – the bifunctional 5,10-methylene-tetrahydrofolate dehydrogenase / 5,10-methylene-tetrahydrofolate cyclohydrolase

(FolD),

the

glycine

cleavage

system

(GcvP,

GcvT,

Lpd),

and

serine

hydroxymethyltransferase (GlyA) – endogenously operate in the reverse direction to that required for formate assimilation. This results in allosteric regulations that might be counterproductive for formate assimilation. For example, FolD is reported to be allosterically inhibited by 10-formyl-THF – the enzyme product under normal conditions – that serves as its substrate in the synthetic pathways.30 Here, we present significant steps towards establishing the reductive glycine pathway and the serine‒ threonine cycle in E. coli. Using E. coli strains auxotrophic to the C1-building blocks and serine, we demonstrate efficient assimilation of formate into these essential cellular components. Glycine is provided either in the medium or is synthesized in vivo via threonine biosynthesis and degradation, according to the expected flux via the serine−threonine cycle. Our results confirm the metabolic feasibility of formate assimilation in E. coli. Based on these results, in the second part of the study we show that we can re-wire the metabolic network of E. coli such that formate becomes an essential intermediate, endogenously produced and consumed and thus connecting two artificially separated parts of the metabolic network. Specifically, we show that under anaerobic conditions, a strain deleted in the natural serine-glycine biosynthesis route can grow on glucose as sole carbon source and satisfies its glycine requirement via threonine degradation and its C1 requirement with endogenously produced formate via the activity of pyruvate formate-lyase. Such a completely prototrophic strain operates via a uniquely re-structured central metabolism that is unprecedented in nature, and could serve as a platform for future metabolicengineering efforts.


Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Results Formate assimilation relieves C1-auxotrophy One-carbon units, carried by THF, are essential for the biosynthesis of purines, thymidine, coenzyme A, and methionine.31 These C1 units provide 2−3% of the carbons required for biomass formation.32 To test whether formate assimilation can satisfy the cellular requirements for C1 units (Figure 2A), we constructed a strain – C1GAUX – that is deleted in serine hydroxymethyltransferase and the enzymes of the glycine cleavage system (∆glyA ∆gcvTHP). As its natural pathways for glycine and one-carbon biosynthesis are blocked (Figure 2A), the C1GAUX strain could not grow on minimal medium with glucose as a sole carbon source. Growth of this strain was possible only when the medium was supplied with glycine as well as intermediates that bypass the need for C1 units, i.e., inosine, thymidine, methionine, and pantothenate (i.e., C1-mix, see Methods), as shown in Figure 2B (dashed line, marked as ‘control’). Upon overexpression of formate THF ligase (FTL) from the methylotrophic bacterium Methylobacterium extorquens33 in the C1GAUX strain, we were able to substitute the C1-mix with formate, hence demonstrating that formate can serve as sole source of the cellular one-carbons (doubling time of 2.3 hours, Figure 2B). To confirm that the C1-dependent building blocks are indeed produced from formate, we fed the bacterium with

13

C-labeled formate and measured the labeling of proteinogenic methionine and

histidine (Methods and Table S1). As methionine’s methyl group is derived from methyl-THF, we expect to see all methionines labeled once. Similarly, as one of histidine's carbons is derived from adenine's sixmembered ring, which in turn originates from formyl-THF, we expect to see all cellular histidines labeled once. Indeed, as shown in Figure 2C, upon feeding with labeled formate both methionine and histidine were almost entirely labeled once, while serine (a non-C1-derived metabolite in this strain) was completely unlabeled. These results confirm that all the C1-moieties carried by THF originated from formate. Formate assimilation relieves serine-auxotrophy Next, we aimed to further extend formate utilization for the synthesis of serine. We constructed a strain – C1SAUX – that is deleted in the first enzyme of the serine biosynthesis pathway as well as in the enzymes of the glycine cleavage system (∆serA ∆gcvTHP). This strain was able to grow on a minimal medium only if supplemented with serine, which can further be metabolized to glycine and 5,10-methylene-THF via serine hydroxymethyltransferase. We tested whether overexpression of FTL enables the strain to replace serine with glycine and formate (Figure 3A). However, we failed to achieve growth (Figure 3B). We speculated that the activity of the enzymes leading from 10-formyl-THF to serine might be limiting. Specifically, it is known that the activity of FolD is allosterically inhibited by 10-formyl-THF, which under normal conditions is the product, rather than the substrate, of this enzyme.30 Hence, we decided to overexpress the endogenous FolD and GlyA (glycine hydroxymethyltransferase). As shown in Figure 3B, overexpression of FolD, alongside FTL, was sufficient to replace serine in the medium with glycine and formate, albeit resulting in retarded growth (doubling time of ~4 hours). However, when FTL, FolD, and



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 28

GlyA were overexpressed together, highly robust growth was achieved under the same conditions (doubling time of 1.6 hours). To confirm that serine was indeed derived from assimilated formate, we analyzed its labeling after feeding the bacterium with

13

C-labeled formate (Table S1). As shown in Figure

3C, virtually all proteinogenic serine molecules (as well as methionine and histidine) were labeled once, as expected. By establishing robust flux from formate to serine we demonstrated the in vivo activity of the upstream segment of both the reductive glycine pathway and the serine−threonine cycle. Threonine biosynthesis and degradation relieves glycine-auxotrophy in a formate-dependent strain Within the serine−threonine cycle, glycine should be recycled via threonine biosynthesis and degradation. We tested whether we can establish threonine degradation as sole source of glycine (and serine) within the C1SAUX strain (Figure 4A). We anticipated that the overexpression of threonine dehydrogenase (Tdh) and 2-amino-3-ketobutyrate CoA ligase (Kbl), enzymes that together cleave threonine to glycine and acetyl-CoA, would suffice to draw high flux towards threonine biosynthesis and glycine production:34 since threonine biosynthesis is highly regulated, a decrease in its concentration (via its degradation) is expected to be compensated by increased biosynthesis.35 Indeed, upon overexpression of Tdh and Kbl, alongside FTL, FolD, and GlyA, we observed growth of the C1SAUX strain on glucose and formate, without the need for exogenously supplied glycine (doubling time of 1.8 hours, Figure 4B). The labeling pattern of serine, methionine, and histidine confirms that formate still provides all the cellular C1 units, including the betacarbon of serine (Figure 4C, Table S1). These results confirm that the enzymes of the serine−threonine cycle can operate together efficiently, as the observed growth depends on high flux from pyruvate/PEP to threonine, from threonine to glycine, and from glycine and formate to serine (which can then be deaminated to pyruvate, thus closing the cycle; see Figure 1B). Re-wiring central metabolism via endogenous production and consumption of formate Apart from a biotechnological relevance, the engineering of formate assimilation in E. coli provides us with a valuable opportunity to test the plasticity of central metabolism and our ability to mold it into novel architectures. Specifically, we were interested in rendering the ∆serA ∆gcvTHP strain prototrophic, using glucose as sole carbon source, by re-wiring its central metabolism. As E. coli is known to produce large amounts of formate under anaerobic conditions via the activity of pyruvate formate-lyase (PflB),36 we expected that under these conditions the addition of formate to the medium will be redundant (Figure 5A). Indeed, when cultivated under anaerobic condition, the C1SAUX strain, overexpressing FTL, FolD, and GlyA, could grow without the addition of formate (doubling time of 1.6 hours, Figure 5B). To confirm that indeed such growth is dependent on endogenously produced and consumed formate, we fed E. coli with glucose labeled in its first, second, or third carbon (alongside


Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


glycine). As pyruvate's carboxylic group – released as formate by pyruvate formate-lyase – is derived from glucose's third carbon, we expected to observe (singly) labeled serine only when feeding with glucose-3-13C, but not with glucose-1-13C or glucose-2-13C. Furthermore, only half of the serine molecules are expected to be labeled as only half of the pyruvate molecules contain the labeled carbon from glucose. As shown in Figure 5C, the labeling of serine matched exactly the anticipated pattern, confirming that pyruvate-derived formate supports C1 and serine biosynthesis in this strain. Next, we tested whether further overexpression of Tdh and Kbl can render the addition of glycine redundant as well, thus establishing growth on glucose as a sole carbon source (Figure 6A). To this aim, we cultivated the C1SAUX strain overexpressing FTL, FolD, GlyA, Tdh, and Kbl under anaerobic conditions. As shown in figure 6B, the resulting strain could sustain growth on glucose as sole carbon source, albeit with a low growth rate and yield (doubling time ~10 hours). We note that unlike growth under aerobic conditions, in which glycine production did not limit growth (Figure 6B), under anaerobic conditions, the requirement of threonine-dependent glycine biosynthesis constrains growth considerably, probably due to a lower flux towards threonine biosynthesis. However, under anaerobic conditions residual growth on glucose was observed even without Tdh and Kbl overexpression, which could be attributed to a higher basal expression of these enzymes. To confirm that the re-wired metabolism is active as speculated, we fed the strain with glucose-1-13C, glucose-2-13C, or glucose-3-13C. The results confirmed the expected fluxes (Figure 6C, D): (1) Feeding with glucose-1-13C resulted in (half of the) pyruvate labeled in the methyl group. Hence formate, produced from PFL, is unlabeled. The oxaloacetate produced from pyruvate's carboxylation is labeled in its methylene group, and hence threonine cleavage results in labeled acetyl-CoA and unlabeled glycine. As a consequence, both glycine and serine (produced from glycine and formate) are fully unlabeled. (2) Feeding with glucose-2-13C results in (half of the) pyruvate labeled in the carbonyl group. Again, formate, produced from PFL, is unlabeled. However, as oxaloacetate is labeled in the carbonyl group, threonine cleavage results in labeled glycine. As a consequence, both glycine and serine are 50% labeled. (3) Finally, feeding with glucose-3-13C results in (half of the) pyruvate labeled in the carboxyl group. Therefore, PFL activity releases a labeled formate. As is the case above, glycine is 50% labeled since oxaloacetate (and threonine) is labeled in the carboxyl group adjacent to the carboxyl (and amine). As serine is synthesized from 50% labeled formate and 50% labeled glycine, serine labeling follows the general trend 1:2:1 for doubly labeled, singly labeled, and unlabeled, respectively.

Discussion To successfully tackle metabolic engineering challenges, it is not enough to restrict oneself to solutions known to operate in nature: as natural pathways are constrained by evolutionary and physiological history, they seldom represent optimal metabolic structures, especially in terms of biotechnological applications37. Formate assimilation exemplifies this quite well. While multiple metabolic pathways are known to support natural formate assimilation, and these are surely suitable for the organisms that use



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 28

them in their habitats, these pathways and organisms are challenging when it comes to biotechnological use (extensively reviewed in7, 9). Generally speaking, there are two strategies that go beyond the limitation of existing pathways37. A ‘mix and match’ approach considers only reactions that are catalyzed by known enzymes. Yet, these enzymes are combined in such a way as to offer a novel pathway that is not known to occur naturally. Another alternative is to evolve enzymes to catalyze novel reactions that could support new routes. While this approach has the advantage of offering countless different candidate pathways to sustain a given metabolic task37, it is constrained by the difficulty of evolving novel reactions at sufficient rates. For example, a recent study put forward a short and ATP-efficient pathway for formate assimilation in which formate is first reduced to formaldehyde, which is then condensed by a newly evolved formolase enzyme to give dihydroxyacetone. However, as the rate of the formolase reaction is very low (kcat/KM < 10 M-1 s-1, four orders of magnitude slower than the average enzyme38), it is difficult to imagine it having a physiological relevance39-40. It therefore seems that that the ‘mix and match’ approach is a more reliable strategy, at least in the short-to-medium

term.

Apart

from

having

advantageous

characteristics

by

itself

–

e.g.,

high

thermodynamic driving force, fast kinetics – a truly promising pathway is one that fits the endogenous metabolism of the host. This usually means trying to avoid imposing non-natural and conflicting fluxes. In previous analyses, we put forward several 'mix and match' pathways that are expected to support efficient formate assimilation in E. coli.7-9 Two of them seem especially interesting (Figure 1). One is the reductive glycine pathway, representing the most efficient aerobic route for formate assimilation. The other, the serine−threonine cycle, is a variant of the naturally occurring serine pathway, which was adapted to the central metabolism of model microbes such as E. coli, bypassing a potentially deleterious flux from malate to acetyl-CoA and glyoxylate, as well as the formation of the toxic intermediate hydroxypyruvate. In this study we presented significant steps towards the realization of both of these synthetic formateassimilation pathways. We took a modular engineering approach for the implementation of the novel pathways, consisting of several steps: (1) Divide the pathway into metabolic modules, each consisting of a set of several consecutive reactions. For example, FTL, FolD, and GlyA constitute the 'THF module' of both pathways; and threonine biosynthesis and degradation correspond to the 'threonine module' of the serine−threonine pathway. (2) Generate deletion strains that are auxotrophic to a set of essential cellular building blocks, and thus require the activity of one of the metabolic modules in order to produce them. For example, the C1SAUX strain (∆serA ∆gcvTHP) is auxotrophic for serine and all cellular building blocks that require C1 (e.g., methionine, purine). (3) Express the enzymes of a metabolic module within the appropriate deletion strain, select for growth – confirming activity of the module – and characterize growth phenotype to identify how it might be possible to enhance activity via further genetic engineering. For example, we were able to confirm high activity of


Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the 'THF module' within the C1SAUX strain only if all module enzymes were overexpressed; expression of FTL and FolD, without GlyA, resulted in retarded growth. In another example, by overexpressing only two enzymes of the 'threonine module', and relying on the endogenous activity of the others, we were able to drive threonine biosynthesis and degradation sufficiently fast to satisfy all cellular needs of glycine and serine (as shown before in a ∆glyA strain34). (4) Selection for module activity can take place in multiple deletion strains, each requiring a different level of activity of the module's enzymes, thus allowing to pinpoint possible bottlenecks that arise only at high flux. For example, in the C1GAUX strain (∆glyA ∆gcvTHP) endogenous activity of FolD was sufficient to sustain growth; however, growth of the C1SAUX strain – for which higher FolD activity is required – necessitated dedicated overexpression of this enzyme. Specifically, since endogenous C1 metabolism in E. coli carries flux in the oxidative direction, FolD evolved to become product-inhibited.30 In our pathways, however, 10-formyl-THF serves as the substrate of FolD and hence its inhibition is deleterious. While we were able to overcome this allosteric inhibition by overexpressing the native FolD, another alternative, for future studies, would be to express a foreign FolD analog that was evolved to carry flux in the reductive direction (e.g., Fch and MtdA from M. extorquens41). (5) Integrate metabolic modules by selecting for their activity in deletion strains that require their combined activity for growth under specific conditions. For example, we integrated the 'THF module' and 'threonine module’ by overexpressing FTL, FolD, GlyA, Kbl, and Tdh within the C1SAUX strain and selecting for growth on glucose and formate. This engineering approach could be highly useful for multiple metabolic-engineering projects, as it dissects novel pathways into manageable units and thus enables direct identification of pathway bottlenecks and testing of possible solutions. Figure 7 summarizes our modular metabolic-engineering approach. Our results confirm that the upstream part of both the reductive glycine pathway and the serine−threonine cycle can operate efficiently in E. coli and provide a substantial portion of its biomass. The combination of glycine synthesis via threonine and serine synthesis from glycine and formate corresponds to the simultaneous activity of all metabolic segments of the serine−threonine cycle, and could thus directly lead to selection for growth on formate via this synthetic pathway. However, several challenges still remain to achieve this goal. First, the production of glycine from threonine seems to be limiting growth, at least under anaerobic conditions; hence, overexpression of threonine biosynthetic enzymes might be beneficial. Also, current flux via the pathway enzymes, even after enhancing glycine production, might not suffice to generate acetyl-CoA at a high enough rate to support growth. Finally, the cycle considerably overlaps with central metabolism, which can result in flux 'leakage' and therefore collapse of pathway activity. Downregulation of enzymes that drain flux from the pathway might therefore be vital (as was demonstrated in42). We showed that under anaerobic conditions, formate produced endogenously via the activity of pyruvate formate-lyase can make redundant the exogenous supply of this compound. When combined with glycine synthesis from threonine, we demonstrated prototrophic growth of the C1SAUX strain, using glucose as ACS Paragon Plus Environment 9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 28

the sole carbon source. This strain represents a substantial re-wiring of central metabolism of E. coli (Figure 8). The construction and characterization of such re-wired strains of model microbes could serve to gain insight into the evolutionary processes that led to the emergence of central metabolism, specifically addressing the question whether its structure is the result of a 'frozen' evolutionary accident or rather represents an optimal solution given a set of biochemical constraints.43 Indeed, the relatively low growth rate and yield of our re-wired strain (Figure 6B) might be improved via long-term evolution, but might also represent an inherent limitation of such a metabolic architecture. From a biotechnological perspective, re-wiring of a host metabolism could optimize bioprocesses to support high efficiency of substrate conversion into product.44 The re-wired prototrophic strain we established in this work is directly relevant for such efforts, as it suggests that even considerable modification of central metabolism could result in viable growth with a direct biotechnological applicability. It would be interesting for future metabolic-engineering efforts to further push the boundaries, establishing metabolic architectures that differ from the canonical central metabolism in any possible aspect.

Materials and Methods Reagents Primers were synthesized by Eurofins (Ebersberg, Germany). PCR reactions were performed using PrimeSTAR GXL DNA Polymerase (Clontech, Heidelberg, Germany). Restriction was carried out using FastDigest enzymes and ligations using T4 DNA ligase, all purchased from Thermo Fisher Scientific (Dreieich,

Germany).

L-glycine,

L-serine,

L-methionine,

inosine,

thymidine,

D-pantothenic

acid

hemicalcium, potassium formate, sodium formate-13C, D-glucose-1-13C and D-glucose-2-13C were ordered from Sigma-Aldrich (Steinheim, Germany). D-glucose was ordered from Carl Roth (Karlsruhe, Germany). D-glucose-3-13C was ordered from Cambridge Isotope Laboratories, Inc. (Andover, MA, USA). Synthetic-operon construction Cloning procedures were designed using Geneious 8 (Biomatters, New Zealand). All overexpressed genes were PCR-amplified from E. coli MG1655 chromosomal DNA: serine hydroxymethyltransferase (glyA), serine deaminase (sdaA), bifunctional 5,10-methylene-tetrahydrofolate dehydrogenase / 5,10-methylenetetrahydrofolate cyclohydrolase (folD), threonine dehydrogenase (tdh), and 2-amino-3-ketobutyrate CoA ligase (kbl). Formate-tetrahydrofolate ligase was synthesized by Thermo Fisher following codon optimization (JCat45) of the ftfL gene of M. extorquens AM1 (GeneBank AY279316.1). The genes were then constructed

into

synthetic

operons,

modulated

each

by

the

ribosome

binding

site

'C'

46

(AAGTTAAGAGGCAAGA), using the 'no background' cloning technique.

The expression plasmid is based on the modular pZ vector 'pZA21MCS' (p15A origin of replication).47 The promoter of this vector was modified to have a constitutive, metabolically relevant, expression level – a mutated form of the phospho-glucose isomerase (pgi) promoter, reported to be ten times stronger than the original.48 The modified plasmid was generated via PCR with the pZpgi primer (GTGGA ATTAT AGCAT TTTAG CCTTT AATTG TCAAT AGGTC TCGAG GTGAA GACGA AAGGG CCTCG) containing the new promoter, ACS Paragon Plus Environment 10

Page 11 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


overriding the sequence of the old one. The kanamycin antibiotic resistance was changed to streptomycin to avoid redundancy with strain deletions. The synthetic operons were inserted to the vectors described above using the restriction enzymes EcoRI and PstI. Bacterial strains E. coli strain DH5α (F- endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG Φ80dlacZ∆M15 ∆(lacZYAargF)U169, hsdR17(rK- mK+), λ-) was used for all cloning procedures. The E. coli strain MG1655 (F- λ- ilvGrfb-50 rph-1) was used for all growth experiments. The entire GCV operon (∆gcvTHP::kan) was deleted in MG1655 by the Red/ET method, recombining the selectable kanamycin resistance instead of the operon (Quick & Easy E. coli Gene Deletion Kit, Gene Bridge, Heidelberg, Germany). Kanamycin resistance was generated via PCR with the primers gcv_KO_F (TAATT TCACC ATGAA AAAGT TGTCA GCCCC GCTTA TTCAA TGAGG ACAAG AATTA ACCCT CACTA AAGGG CG) and gcv_KO_R (CTGAC TAAAA AGGCG CCGAA GCGCC TTTAG AAAAT AGTCG AATCA GTGAA TAATA CGACT CACTA TAGGG CTC), with the plasmid pKD3 as template (GeneBank: AY048742). The kanamycin cassette – flanked by FRT sites that allow its removal by the FLP mechanism49 – was recombined into the genome. Keio collection50 deletion strains JW2535 (∆glyA::kan) and JW2880 (∆serA::kan) were used to transfer their gene deletion to the ∆gcvTHP strain via P1 phage transduction.51 The resulting strains were annotated as C1GAUX (∆gcvTHP, ∆glyA::kan), and C1SAUX (∆gcvTHP, ∆serA::kan), respectively. Growth conditions LB medium (1% NaCl, 0.5% yeast extract, and 1% tryptone) was used for strain propagation (preexperiment). Growth of the C1GAUX strain required the addition of 1 mM glycine and 0.3 mM thymidine, whereas overexpression of FTL rendered these supplements unnecessary. Growth experiments were performed in M9 minimal media (50 mM Na2HPO4, 20 mM KH2PO4, 1 mM NaCl, 20 mM NH4Cl, 2 mM MgSO4 and 100 µM CaCl2), supplemented with trace elements (134 µM EDTA, 13 µM FeCl3∙6H2O, 6.2 µM ZnCl2, 0.76 µM CuCl2∙2H2O, 0.42 µM CoCl2∙2H2O, 1.62 µM H3BO3, 0.081 µM MnCl2∙4H2O). Carbon sources were added according to strain and specific experiment: 10 mM glucose, 1 mM glycine, 1 mM serine, 5 mM formate, and C1-mix (0.3 mM methionine, 0.3 mM thymidine, 0.3 mM inosine, 6 µM pantothenate). Pre-cultures were cultivated in 3 mL of M9 medium containing either (i) glucose, glycine, and C1-mix (C1GAUX); (ii) glucose, serine (C1SAux); (iii) glucose, glycine, and formate (for all strains expressing FTL). Pre-cultures were grown at 37ºC, 220−250 RPM. Prior to inoculation cells were harvested by centrifugation (11,000 RPM, 30 sec, 4ºC), washed three times in ice-cold M9, and inoculated in supplemented M9 media to an optical density of 0.005. Nunc 96-well microplates were used for growth experiments (Thermo Fisher Scientific), each well containing 150 µl culture covered with 50 µl mineral oil (Sigma-Aldrich) to avoid evaporation. For growth experiments under anaerobic conditions, M9 medium was placed for at least 24 h in the anaerobic chamber to allow removal of dissolved oxygen from the media.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 28

Growth experiments were carried out in a Tecan Infinite 200 Pro plate reader (Tecan, Switzerland) at 37ºC. Aerobically, growth (OD600) was measured after a kinetic cycle of 12 shaking steps, which alternate between linear and orbital (1 mm amplitude), each 60 sec long. For anaerobic conditions, the plate reader was stationed in a vinyl chamber (N2 with 10% CO2, 2.5% H2, model B, Coy Laboratory Products, Grass Lake, MI, USA). Growth (OD600) was measured after a kinetic cycle consisting of 14.5 min without shaking, 60 sec orbital shaking (6 mm amplitude), further 20 sec without shaking, and measurement of absorbance at 600 nm. OD values measured in the plate reader were calibrated to represent

OD

values

of

standard

cuvette,

which

were

found

to

correlate

according

to

ODcuvette=ODplate/0.23. All growth experiments were performed in triplicates, and the growth curves shown represent the average of these triplicates (in all cases variability between triplicates was < 5%). Carbon labeling For stationary isotope tracing of proteinogenic amino acids,52 cells were cultured in 3 mL M9 media under conditions as described above, and were supplemented with either labeled or unlabeled carbon sources, e.g., formate-13C and glucose-1-13C, glucose-2-13C, glucose-3-13C. After reaching stationary phase, ~109 cells were harvested by centrifugation for 1 min at 11,000 RPM. The biomass was hydrolyzed by incubation with 1 mL 6N hydrochloric acid for a duration of 24 h in 95°C. The acid was then evaporated by continued heating at 95°C and nitrogen streaming. Hydrolyzed amino acids were separated using ultra-performance liquid chromatography (Acquity, Waters, Milford, MA, USA) using a C18-reversed-phase column (Waters) as previously described.53 Mass spectra were acquired using an Exactive mass spectrometer (Thermo Fisher). Data analysis was performed using Xcalibur (Thermo Fisher). Prior to analysis, authentic amino-acid standards (Sigma-Aldrich) were analyzed under the same conditions in order to determine typical retention times. The results of all labeling experiments are shown in Table S1.


Page 13 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Supporting information The Supporting Information is available free of charge on the ACS Publications website at DOI: XXXXXX Labeling pattern of proteinogenic amino-acids for all stains and conditions analyzed in this study.

ORCID Orens Yishai: 0000-0001-8712-3959 Hezi Tenenboim: 0000-0002-3783-9446 Steffen N. Lindner: 0000-0001-6910-6573 Arren Bar-Even: 0000-0002-1039-4328

Acknowledgements The authors thank Änne Michaelis and Lothar Willmitzer for vital help with the metabolomic analysis and Anika Bauer, Jost Lühle, and Sarah Amina Ahrendt for technical assistance. The authors further thank Madeleine Bouzon-Bloch, Volker Doring, Philippe Marliere, Charles A. R. Cotton, and Niv Antonovsky for helpful discussions and suggestions. The work was funded by the Max Planck Society and by the German Ministry of Education and Research (project ‘FormatPlant’, part of BioEconomy 2030, Plant Breeding Research for the Bioeconomy).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 28

References (1) Burg, J. M.; Cooper, C. B.; Ye, Z.; Reed, B. R.; Moreb, E. A.; Lynch, M. D. (2016) Large-scale bioprocess competitiveness: the potential of dynamic metabolic control in two-stage fermentations, Curr Opin Chem Eng. 14, 121-136. (2) Burk, M. J.; Van Dien, S. (2016) Biotechnology for chemical production: challenges and opportunities, Trends Biotechnol 34, 187-90. (3) Tollefson, J. (2008) Energy: not your father's biofuels, Nature 451, 880-3. (4) Naik, S. N.; Goud, V. V.; Rout, P. K.; Dalai, A. K. (2010) Production of first and second generation biofuels: a comprehensive review, Renew. Sustainable Energy Rev. 14, 578-597. (5) Chundawat, S. P.; Beckham, G. T.; Himmel, M. E.; Dale, B. E. (2011) Deconstruction of lignocellulosic biomass to fuels and chemicals, Annual review of chemical and biomolecular engineering 2, 121-45. (6) Carriquiry, M. A.; Du, X.; Timilsina, G. R. (2011) Second generation biofuels: Economics and policies, Energy Policy 39, 4222-34. (7) Yishai, O.; Lindner, S. N.; Gonzalez de la Cruz, J.; Tenenboim, H.; Bar-Even, A. (2016) The formate bio-economy, Curr Opin Chem Biol 35, 1-9. (8) Bar-Even, A.; Noor, E.; Flamholz, A.; Milo, R. (2013) Design and analysis of metabolic pathways supporting formatotrophic growth for electricity-dependent cultivation of microbes, Biochim Biophys Acta 1827, 1039-47. (9) Bar-Even, A. (2016) Formate Assimilation: The metabolic architecture of natural and synthetic pathways, Biochemistry 55, 3851-63. (10) Agarwal, A. S.; Zhai, Y.; Hill, D.; Sridhar, N. (2011) The electrochemical reduction of carbon dioxide to formate/formic acid: engineering and economic feasibility, ChemSusChem 4, 1301-10. (11) Kopljar, D.; Inan, A.; Vindayer, P.; Wagner, N.; Klemm, E. (2014) Electrochemical reduction of CO2 to formate at high current density using gas diffusion electrodes, J. Appl. Electrochem. 44, 1107-1116. (12) Wang, Q.; Dong, H.; Yu, H. (2014) Fabrication of a novel tin gas diffusion electrode for electrochemical reduction of carbon dioxide to formic acid, RSC Adv 4, 59970-59976. (13) Lu, X.; Leung, D. Y.; Wang, H.; Leung, M. K.; Xuan, J. (2014) Electrochemical reduction of carbon dioxide to formic acid, ChemElectroChem 1, 836-849. (14) Pletcher, D. (2015) The cathodic reduction of carbon dioxide—What can it realistically achieve? A mini review, Electrochem Commun 61, 97-101. (15) Taheri, A.; Berben, L. A. (2016) Making C-H bonds with CO2: production of formate by molecular electrocatalysts, Chem Commun (Camb) 52, 1768-77. (16) Tamaki, Y.; Morimoto, T.; Koike, K.; Ishitani, O. (2012) Photocatalytic CO2 reduction with high turnover frequency and selectivity of formic acid formation using Ru(II) multinuclear complexes, Proc Natl Acad Sci U S A 109, 15673-8. (17) Wang, W. H.; Himeda, Y.; Muckerman, J. T.; Manbeck, G. F.; Fujita, E. (2015) Hydrogenation to formate and methanol as an alternative to photo-and electrochemical CO2 reduction, Chem. Rev. 115, 12936−12973. (18) Enthaler, S. (2008) Carbon dioxide—The hydrogen-storage material of the future?, ChemSusChem 1, 801-804. (19) Wölfel, R.; Taccardi, N.; Bösmann, A.; Wasserscheid, P. (2011) Selective catalytic conversion of biobased carbohydrates to formic acid using molecular oxygen, Green Chem. 13, 2759-2763. (20) Albert, J.; Lüders, D.; Bösmann, A.; Guldi, D. M.; Wasserscheid, P. (2014) Spectroscopic and electrochemical characterization of heteropoly acids for their optimized application in selective biomass oxidation to formic acid, Green Chemistry 16, 226-237. (21) Albert, J.; Wölfel, R.; Bösmann, A.; Wasserscheid, P. (2012) Selective oxidation of complex, waterinsoluble biomass to formic acid using additives as reaction accelerators, Energy Environ Sci 5, 79567962. (22) Sorokin, A. B.; Kudrik, E. V.; Alvarez, L. X.; Afanasiev, P.; Millet, J. M. M.; Bouchu, D. (2010) Oxidation of methane and ethylene in water at ambient conditions, Catal. Today 157, 149-154. (23) Shukla, R. S.; Bhatt, S. D.; Thorat, R. B.; Jasra, R. V. (2005) A novel effective hydration of carbon monoxide in liquid phase by a water-soluble ruthenium complex catalyst at moderate pressures in aqueous medium, ppl. Catal., A 294, 111-118. (24) Grunwald, S.; Mottet, A.; Grousseau, E.; Plassmeier, J. K.; Popovic, M. K.; Uribelarrea, J. L.; Gorret, N.; Guillouet, S. E.; Sinskey, A. (2015) Kinetic and stoichiometric characterization of organoautotrophic growth of Ralstonia eutropha on formic acid in fed-batch and continuous cultures, Microb Biotechnol 8, 155-63. ACS Paragon Plus Environment 14

Page 15 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(25) Schiel-Bengelsdorf, B.; Durre, P. (2012) Pathway engineering and synthetic biology using acetogens, FEBS Lett 586, 2191-8. (26) Shen, Y.; Jarboe, L.; Brown, R.; Wen, Z. (2015) A thermochemical-biochemical hybrid processing of lignocellulosic biomass for producing fuels and chemicals, Biotechnology advances 33, 1799-813. (27) Schrader, J.; Schilling, M.; Holtmann, D.; Sell, D.; Filho, M. V.; Marx, A.; Vorholt, J. A. (2009) Methanol-based industrial biotechnology: current status and future perspectives of methylotrophic bacteria, Trends Biotechnol 27, 107-15. (28) Kim, J.; Copley, S. D. (2012) Inhibitory cross-talk upon introduction of a new metabolic pathway into an existing metabolic network, Proc Natl Acad Sci U S A 109, E2856-64. (29) Sah, S.; Aluri, S.; Rex, K.; Varshney, U. (2015) One-carbon metabolic pathway rewiring in Escherichia coli reveals an evolutionary advantage of 10-formyltetrahydrofolate synthetase (Fhs) in survival under hypoxia, J Bacteriol 197, 717-26. (30) Dev, I. K.; Harvey, R. J. (1978) A complex of N5,N10-methylenetetrahydrofolate dehydrogenase and N5,N10-methenyltetrahydrofolate cyclohydrolase in Escherichia coli. Purification, subunit structure, and allosteric inhibition by N10-formyltetrahydrofolate, J Biol Chem 253, 4245-53. (31) Maden, B. E. (2000) Tetrahydrofolate and tetrahydromethanopterin compared: functionally distinct carriers in C1 metabolism, Biochem J 350 Pt 3, 609-29. (32) Neidhardt, F. C.; Ingraham, J. L.; Schaechter, M. (1990) Building blocks needed to produce 1g of E. oli protoplasm, In Physiology of the Bacterial Cell: A Molecular Approach, pp 134-143. (33) Marx, C. J.; Laukel, M.; Vorholt, J. A.; Lidstrom, M. E. (2003) Purification of the formatetetrahydrofolate ligase from Methylobacterium extorquens AM1 and demonstration of its requirement for methylotrophic growth, J Bacteriol 185, 7169-75. (34) Fraser, J.; Newman, E. B. (1975) Derivation of glycine from threonine in Escherichia coli K-12 mutants, J Bacteriol 122, 810-7. (35) Yuzbashev, T. V.; Vybornaya, T. V.; Larina, A. S.; Gvilava, I. T.; Voyushina, N. E.; Mokrova, S. S.; Yuzbasheva, E. Y.; Manukhov, I. V.; Sineoky, S. P.; Debabov, V. G. (2013) Directed modification of Escherichia coli metabolism for the design of threonine-producing strains, Appl. Biochem. Microbiol. 49, 723-742. (36) Knappe, J.; Sawers, G. (1990) A radical-chemical route to acetyl-CoA: the anaerobically induced pyruvate formate-lyase system of Escherichia coli, FEMS Microbiol Rev 6, 383-98. (37) Erb, T. J.; Jones, P. R.; Bar-Even, A. (2017) Synthetic metabolism: metabolic engineering meets enzyme design, Curr Opin Chem Biol 37, 56-62. (38) Bar-Even, A.; Noor, E.; Savir, Y.; Liebermeister, W.; Davidi, D.; Tawfik, D. S.; Milo, R. (2011) The moderately efficient enzyme: evolutionary and physicochemical trends shaping enzyme parameters, Biochemistry 50, 4402-10. (39) Siegel, J. B.; Smith, A. L.; Poust, S.; Wargacki, A. J.; Bar-Even, A.; Louw, C.; Shen, B. W.; Eiben, C. B.; Tran, H. M.; Noor, E.; Gallaher, J. L.; Bale, J.; Yoshikuni, Y.; Gelb, M. H.; Keasling, J. D.; Stoddard, B. L.; Lidstrom, M. E.; Baker, D. (2015) Computational protein design enables a novel one-carbon assimilation pathway, Proc Natl Acad Sci U S A 112, 3704-9. (40) Poust, S.; Piety, J.; Bar-Even, A.; Louw, C.; Baker, D.; Keasling, J. D.; Siegel, J. B. (2015) Mechanistic analysis of an engineered enzyme that catalyzes the formose reaction, Chembiochem. (41) Crowther, G. J.; Kosaly, G.; Lidstrom, M. E. (2008) Formate as the main branch point for methylotrophic metabolism in Methylobacterium extorquens AM1, J Bacteriol 190, 5057-62. (42) Antonovsky, N.; Gleizer, S.; Noor, E.; Zohar, Y.; Herz, E.; Barenholz, U.; Zelcbuch, L.; Amram, S.; Wides, A.; Tepper, N.; Davidi, D.; Bar-On, Y.; Bareia, T.; Wernick, D. G.; Shani, I.; Malitsky, S.; Jona, G.; Bar-Even, A.; Milo, R. (2016) Sugar Synthesis from CO2 in Escherichia coli, Cell 166, 115-25. (43) Bar-Even, A.; Flamholz, A.; Noor, E.; Milo, R. (2012) Rethinking glycolysis: on the biochemical logic of metabolic pathways, Nat Chem Biol 8, 509-517. (44) Lynch, M. D. (2016) Into new territory: improved microbial synthesis through engineering of the essential metabolic network, Curr Opin Biotechnol 38, 106-11. (45) Grote, A.; Hiller, K.; Scheer, M.; Munch, R.; Nortemann, B.; Hempel, D. C.; Jahn, D. (2005) JCat: a novel tool to adapt codon usage of a target gene to its potential expression host, Nucleic Acids Res 33, W526-31. (46) Zelcbuch, L.; Antonovsky, N.; Bar-Even, A.; Levin-Karp, A.; Barenholz, U.; Dayagi, M.; Liebermeister, W.; Flamholz, A.; Noor, E.; Amram, S.; Brandis, A.; Bareia, T.; Yofe, I.; Jubran, H.; Milo, R. (2013) Spanning high-dimensional expression space using ribosome-binding site combinatorics, Nucleic Acids Res 41, e98. (47) Lutz, R.; Bujard, H. (1997) Independent and tight regulation of transcriptional units in Escherichia coli via the LacR/O, the TetR/O and AraC/I1-I2 regulatory elements, Nucleic Acids Res 25, 1203-10. ACS Paragon Plus Environment 15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 28

(48) Braatsch, S.; Helmark, S.; Kranz, H.; Koebmann, B.; Jensen, P. R. (2008) Escherichia coli strains with promoter libraries constructed by Red/ET recombination pave the way for transcriptional fine-tuning, Biotechniques 45, 335-7. (49) Datsenko, K. A.; Wanner, B. L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products, Proc Natl Acad Sci U S A 97, 6640-5. (50) Baba, T.; Ara, T.; Hasegawa, M.; Takai, Y.; Okumura, Y.; Baba, M.; Datsenko, K. A.; Tomita, M.; Wanner, B. L.; Mori, H. (2006) Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection, Mol Syst Biol 2, 2006-2008. (51) Thomason, L. C.; Costantino, N.; Court, D. L. (2007) E. coli genome manipulation by P1 transduction, Curr Protoc Mol Biol Chapter 1, Unit 1 17. (52) Thiele, B.; Stein, N.; Oldiges, M.; Hofmann, D. (2012) Direct analysis of underivatized amino acids in plant extracts by LC-MS-MS, Methods Mol Biol 828, 317-28. (53) Giavalisco, P.; Li, Y.; Matthes, A.; Eckhardt, A.; Hubberten, H. M.; Hesse, H.; Segu, S.; Hummel, J.; Kohl, K.; Willmitzer, L. (2011) Elemental formula annotation of polar and lipophilic metabolites using (13) C, (15) N and (34) S isotope labelling, in combination with high-resolution mass spectrometry, Plant J 68, 364-76. (54) de Lorenzo, V.; Sekowska, A.; Danchin, A. (2014) Chemical reactivity drives spatiotemporal organisation of bacterial metabolism, FEMS Microbiol Rev 39, 96-119.


Page 17 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure Legends Figure 1 Two promising synthetic formate assimilation pathways. Red and yellow shadings denote the formate and CO2 input of the pathways, respectively. The products of the pathways are shaded in blue. (A) The reductive glycine pathway, the most efficient route for formate assimilation under aerobic conditions. In this pathway two formate molecules and one CO2 molecule are condensed to form pyruvate at the expense of two ATP and three NAD(P)H molecules. (B) The serine‒threonine pathway, a variant of the methylotrophic serine pathway. One turn of the cycle assimilates one formate molecule and one CO2 molecule into acetyl-CoA. Asd

corresponds

to

aspartate

semialdehyde

dehydrogenase

(1.2.1.11);

AspC

to

aspartate

aminotransferase (EC 2.6.1.1); GcvH to a lipoyl carrier protein; GcvT to aminomethyltransferase (EC 2.1.2.10);

GcvP

to

glycine

dehydrogenase

(decarboxylating)

(EC

1.4.4.2);

GlyA

to

serine

hydroxymethyltransferase (EC 2.1.2.1); FolD to bifunctional methenyltetrahydrofolate cyclohydrolase / dehydrogenase (EC 3.5.4.9 and 1.5.1.5/15); FTL to formate tetrahydrofolate ligase (EC 6.3.4.3); Kbl to 2amino-3-ketobutyrate CoA ligase (EC 2.3.1.29); Lpd to dihydrolipoyl dehydrogenase (EC 1.8.1.4); Ppc to phosphoenolpyruvate carboxylase (E.C 4.1.1.31); Pps corresponds to phosphoenolpyruvate synthetase (EC 2.7.9.2); SdaA to serine deaminase (EC 4.3.1.17); Tdh to threonine dehydrogenase (EC 1.1.1.103); ThrA to aspartate kinase / homoserine dehydrogenase (EC 2.7.2.4 and 1.1.1.3); ThrB to homoserine kinase (EC 2.7.1.39); and ThrC to threonine synthase (EC 4.2.3.1). Figure 2 Formate assimilation via formate-THF ligase relieves C1-auxotrophy. (A) The designed selection scheme in which formate serves as sole source for the biosynthesis of 10-formyl-THF, and 5,10-methylene-THF. Compounds circled in green are carbon sources and deletions are marked in red. Enzyme ‘F’ is FTL and ‘D’ is FolD. (B) Growth of the C1GAUX strain (∆glyA ∆gcvTHP) on different carbon sources, with and without overexpression of FTL. ‘Control’ refers to growth of the above strain, without overexpression, on glucose supplemented with C1-mix, which comprises of all required C1-derived compounds: methionine, thymidine, inosine and pantothenate. (C) Labeling pattern of proteinogenic histidine (‘H’), methionine (‘M’), and serine (‘S’). WT E. coli (MG1655) was fed on glucose, glycine, and formate-13C. C1GAUX strain expressing FTL (C1G-F) was fed with glucose, glycine, and either unlabeled or

13

C-labeled formate.

Figure 3 Formate assimilation via formate-THF ligase relieves serine auxotrophy. (A) The designed selection scheme in which formate serves as an essential source for the biosynthesis of 10-formyl-THF, 5,10methylene-THF, and serine. Compounds circled in green are carbon sources and deletions are marked in red. Enzyme ‘F’ is FTL, ‘D’ is FolD, and ‘G’ is GlyA. (B) Growth of the C1SAUX strain (∆serA ∆gcvTHP) on different carbon sources, with and without overexpression of FTL, FolD, and/or GlyA. ‘Control’ refers to



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 28

growth of the above strain, without overexpression, on glucose supplemented with serine; the growth yield was low probably due to serine toxicity.54 (C) Labeling pattern of proteinogenic histidine (‘H’), methionine (‘M’), and serine (‘S’). WT E. coli (MG1655) was fed on glucose, glycine, and formate-13C. C1GAUX strain expressing FTL, FolD, and GlyA (C1S-FDG) was fed with glucose, glycine, and either unlabeled or

13

C-labeled formate.

Figure 4 Threonine biosynthesis and degradation relieves glycine-auxotrophy in a formate-dependent strain. (A) The designed selection scheme in which formate serves as an essential carbon source and glycine is produced solely from threonine. Compounds circled in green are carbon sources and deletions are marked in red. Enzyme ‘F’ is FTL, ‘D’ is FolD, ‘G’ is GlyA, ‘K’ is Kbl, and ‘T’ is Tdh. (B) Growth of the C1GAUX strain on different carbon sources, with and without overexpression of FTL, FolD, GlyA, Kbl, and/or Tdh. (Growth results of the C1GAUX strain that does not overexpress any enzymes, i.e., ‘Null’, are shown in Figure 3.) (C) Labeling pattern of proteinogenic histidine (‘H’), methionine (‘M’), and serine (‘S’). WT E. coli (MG1655) was fed on glucose and formate-13C. C1GAUX strain expressing FTL, FolD, GlyA, Kbl, and Tdh (C1S-TKFDG) was fed with glucose, and either unlabeled or

13

C-labeled formate.

Figure 5 Formate produced endogenously relieves serine auxotrophy. (A) The designed selection scheme in which formate is produced by pyruvate formate lyase (enzyme ’P’) and serves as an essential carbon source for the production of 10-formyl-THF, 5,10-methylene-THF, and serine. Compounds circled in green are carbon sources and deletions are marked in red. Enzyme ‘F’ is FTL, ‘D’ is FolD, and ‘G’ is GlyA. (B) Growth of the C1GAUX strain on different carbon sources, with and without overexpression of FTL, FolD, and/or GlyA. ‘Control’ refers to growth of the above strain, without overexpression, on glucose supplemented with serine. (C) Labeling pattern of proteinogenic serine – within the C1GAUX strain expressing FTL, FolD, and GlyA (C1S-FDG) – upon feeding with glycine and glucose labeled in its 1st, 2nd, or 3rd position. Figure 6 Endogenous production of glycine and serine enables growth of C1GAUX strain. (A) The designed selection scheme in which formate is produced by pyruvate formate lyase (enzyme ’P’) and glycine is produced from threonine. Compounds circled in green are carbon sources and deletions are marked in red. Enzyme ‘F’ is FTL, ‘D’ is FolD, ‘G’ is GlyA, ‘K’ is Kbl, and ‘T’ is Tdh. (B) Growth of the C1GAUX strain on different carbon sources, with and without overexpression of FTL, FolD, GlyA, Kbl, and/or Tdh. (C) Labeling pattern of proteinogenic serine – within the C1GAUX strain expressing FTL, FolD, GlyA, Kbl, and Tdh (C1S-TKFDG) – upon feeding with glucose labeled in its 1st, 2nd, or 3rd position. (D) Expected labeling in glycine and serine upon feeding with glucose labeled in its 1st, 2nd, or 3rd position. When only half of the molecules are expected to be labeled at a certain position, the corresponding carbon is shown as half-full. The different colors serve only to emphasize the origin of the labeling for the different carbon atoms of serine. We note that as growth took place under fermentative conditions (no electron acceptor), the TCA cycle operates in


Page 19 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


a fork mode, thereby (almost) eliminating any reshuffling of the carbon labeling that would otherwise result from a cyclic flow through the pathway. Figure 7 A schematic representation of the modular metabolic-engineering approach demonstrated in this paper, as described in detail in the text. In this study we demonstrated steps A-D. Figure 8 A completely prototrophic strain in which central metabolism was re-wired, such that serine, glycine, and C1-metabolism proceed via a unique set of pathways. This strain can grow on glucose (circled in green) as sole carbon source under anaerobic conditions.



glucose

glycolysis

PEP

C1 metabolism THF

pyruvate

PFL formate FTL acetyl-CoA

CHO-THF CH2-THF serine

serine glycine metabolism

TCA cycle

acetyl-CoA

threonine

HCO3-

anaplerosis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

oxaloacetate

aspartate

threonine metabolism

homoserine

ACS Paragon Plus Environment

Page 20 of 28

Page 21 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(A)

(B)

formate THF

FTL

THF GcvH GcvT GcvP Lpd

CO2

formate

10-formyl-THF

H2O

glycine

GlyA

THF

acetyl-CoA

serine SdaA

NADPH NADP+, H2O

FolD

5,10-methylene-THF

NADH, NH3 NAD+

ATP ADP, Pi

10-formyl-THF

NADPH NADP+, H2O

FolD

THF

FTL

ATP ADP, Pi

NH3

pyruvate

Kbl

5,10-methylene-THF H2O

glycine CoA

serine SdaA

THF

2-amino3-oxobutyrate Tdh

GlyA

pyruvate

NADH NAD+

Pps

2 NADP+, 2 ADP, 2 Pi 2 NADPH, 2 ATP, H2O AspC

Ppc

threonine ThrA ThrB ThrC Asd

aspartate

ATP, H2O AMP, Pi

PEP CO 2 Pi

oxaloacetate

2-ketoglutarate glutamate


NH3


(A)

1.8

formate F

10-formyl-THF THF

D

5,10-methylene-THF ΔgcvTHP ΔglyA glycine

(B)

(C)

THF

serine

% 100

1.4 OD (600nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 28

TCA cycle / glycolysis

H M S

H M S

75

Overexpression

1.0

Null

F

Carbon Sources control glu+gly+for glu+gly glu+for

0.6

THF

glucose

H M S

0.2 10

20

30 40 Time (hours)

50


50 25 0

WT: C1G-F: C1G-F: glu+gly+13C-for glu+gly+for glu+gly+13C-for # labeled carbons 0 1 2

Page 23 of 28

(A)

formate F

1.8

THF

10-formyl-THF THF

D D

5,10-methylene-THF ΔgcvTHP G glycine

serine

(C)

(B) Overexpression Null

1.4 OD (600nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


F F D F G

1.0

F D G

Carbon Sources control glu+gly+for glu+gly glu+for

0.6

THF

glucose

ΔserA TCA cycle / glycolysis

% 100

0.2 10

20

30 40 Time (hours)

50


H M S

H M S

H M S

75 50 25 0

WT: C1S-FDG: C1S-FDG: glu+gly+13C-for glu+gly+for glu+gly+13C-for # labeled carbons 0 1 2


(A)

1.8

formate F

10-formyl-THF THF

D


T K

(C)

(B)

THF

serine

% 1.4 OD (600nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 28

100

Overexpression

1.0

F D G T K F D G

Carbon Sources glu+gly+for glu+gly glu+for

THF

threonine glucose


0.2 10

20

30 40 Time (hours)

50


H M S

H M S

75 50 25 0

0.6

H M S

C1S-FDGTK: C1S-FDGTK: WT: glu+for glu+13C-for glu+13C-for # labeled carbons 0 1 2

Page 25 of 28

(A)

(B)

formate P

F

(C) 1.0

THF

%

Overexpression

10-formyl-THF THF

D

5,10-methylene-THF ügcvTHP G glycine

OD (600nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Null

F

0.6

F D F G F D G

serine THF

Carbon Sources control glu+gly+for glu+gly

0.2

glucose üserA TCA cycle / glycolysis

10

20

30 40 Time (hours)

50


100

S

S

S

75 50 25 0

13

C1

13

C2

13

C3

13

C1S-FDG: C-glu+gly # labeled carbons 0 1 2


(A)

(B)

formate

P

THF

F

(C) 1.0

%

10-formyl-THF D

THF


OD (600nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 28

0.6


(D) O O

O

glucose

O

O

O

1 2

AcCoA

glycine

1

AcCoA

formate

O

1 2 3 4

O

threonine

20

30 40 50 Time (hours) O

O

serine O

O

10

1 2 3 4 5 6

1 2 3

O

O

pyruvate

O

O

O

O oxaloacetate O

O

1 2

AcCoA

glycine

1

AcCoA

formate

0

13

C1

glucose

serine

pyruvate

O

threonine

O

C3

1 2 3 4

O

O

oxaloacetate O

O


1 2 3 4 5 6

1 2 3

O

O

13

# labeled carbons 0 1 2

O

O

C2

C1S-TKFDG: C-glu

1 2 3

1 2 3 4

13

13

O

O

S G

25

O O

glucose O

1 2 3 4

50

S G

60

serine

1 2 3

S G

1 2 3 4 5 6

1 2 3

O

O

75

Carbon Sources glu+gly+for glu+gly glu+for glu

0.2

threonine glucose

Null

F D G T K F D G

serine

THF

T K

Overexpression

100

1 2

AcCoA

glycine

1

AcCoA

formate

O

1 2 3 4

O

threonine

O

1 2 3

O

pyruvate

O

1 2 3 4

O

O oxaloacetate O

Page 27 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A


B

C

Divide synthetic pathway into several metabolic modules, each corresponds to a defined metabolic task

D

Increase selection pressure for module activity in different strains

E + =

Select for the activity of the entire synthetic pathway (combining all modules) & optimize growth properties

+ + Select each module in an auxotrophic strain, the growth of which depends upon module activity

low selection

Integrate modules and select for high combined activity selection in other strains


=


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 28

glucose glycolysis PEP PykA

anaplerosis ADP ATP

pyruvate C1 metabolism FTL

THF

ATP

ADP, Pi

CoA

PflB

FolD

H2O

5,10-methylene-THF

AspC

CoA

2-amino-3-ketobutyrate Tdh

NADH NAD+

threonine

Pi H2O ThrC

glutamate 2-ketoglutarate

aspartate

TCA cycle

ThrA

ATP ADP

aspartyl phosphate

glycine Acetyl-CoA

Kbl

Pi

oxaloacetate

∆gcvTHP

GlyA

serine/glycine metabolism

NADPH NADP+ H2O

HCO3-

acetyl-CoA

10-formyl-THF

∆serA

serine

formate

Ppc

threonine biosynthesis and cleavage

homoserine phosphate

Asd

NADPH NADP+, Pi

aspartate semialdehyde ADP ATP ThrA ThrB

NADPH NADP+

homoserine


Escherichia coli - ACS Publications - American Chemical Society

Recommend Documents