Carbohydrate metabolism in Fibrobacter succinogenes: What NMR tells us

Abstract

Fibrobacter succinogenes is a major rumen fibrolytic bacterium found in high numbers when ruminants are fed cellulose-rich diets. It produces a very efficient fibrolytic system comprising numerous enzymes, but the organization and catalytic features of this system remain unclear. We were using nuclear magnetic resonance (NMR) to study carbohydrate metabolism by Fibrobacter spp. well before the term ‘metabolomics’ was coined. We first analyzed in detail the NMR spectra arising from resting cells of F. succinogenes provided with glucose, cellobiose or cellulose as a substrate. We were able to show carbohydrate cycling and reversibility of some steps. The use of 1D and 2D ¹H and ¹³C NMR showed the synthesis by the cells of unexpected oligosaccharides: maltodextrins, maltodextrin-1-phosphate, and other glucose derivatives. We also showed that maltodextrins and maltodextrins-1-phosphate were synthesized and excreted out of the cell, and could originate from exogenous sugars or endogenous glycogen. According to the substrate and the physiological state of the bacterium (resting or growing cells), metabolic deviations toward the synthesis of these oligosaccharides were observed. We have also used liquid- and solid-state NMR to monitor wheat straw degradation by F. succinogenes. The originality of this NMR approach was to investigate the activity of an entire fibrolytic system on an intact complex substrate (wheat straw). No preferential degradation of amorphous versus crystalline cellulose was observed, nor of cellulose versus hemicelluloses. We showed for the first time a sequential activity of some enzymes of F. succinogenes S85 on wheat straw. In conclusion, our studies illustrate the utility of various NMR approaches to study and better understand sugar and polysaccharide metabolism in rumen bacteria.

• ruminants
• NMR
• Fibrobacter succinogenes
• polysaccharide metabolism
• cellulose

Introduction

Fibrobacter succinogenes is a predominant cellulolytic rumen bacterium found in high numbers when ruminants are fed cellulose-rich diets. This bacterium was first isolated by Hungate in 1947 [1] and since then its cellulolytic system and its physiology have been extensively studied by several groups. F. succinogenes is particularly efficient in degrading various forms of crystalline cellulose and it shows a high ability to solubilize different plant cell wall polysaccharides [2], [3]. Plant cell walls are degraded by a very efficient fibrolytic system encoded by a high number of genes, as suggested by the genome sequence of strain S85 completed recently, which showed more than 100 genes encoding enzymes potentially active against carbohydrates [3], [4]. However, the organization and catalytic features of this system remain unclear. Cellulose, the sole substrate for F. succinogenes S85 within plant cell wall, is hydrolyzed into cellobiose and glucose that are metabolized by the cells through the Embden-Meyerhof-Parnass (EMP) glycolytic pathway and fermented to succinate, acetate, and a little formate (Figure 1), [5], [6]. It was also shown that F. succinogenes was able to store high amounts of intracellular glycogen [7]. We started to use NMR to revisit the metabolism of this bacterium. NMR is a tool of choice to describe microbial metabolic pathways [8]. It relies on the use of enriched precursors for improved sensitivity and selectivity, the recording of successive spectra in situ without destruction of the cells to monitor the metabolic route of a specific atom, and examination of cell extracts for precise identification of metabolites. We employed ¹³C and ¹H NMR spectroscopy to study [1-¹³C]glucose metabolism in F. succinogenes S85, using both living cells (‘resting cells’) and acellular extracts. We showed interesting features such as glycogen cycling and reversibility of some pathways in this strain, but also in other strains of the genus. The use of 1D and 2D ¹H and ¹³C NMR showed the synthesis by the cells of unexpected oligosaccharides, whose origin was investigated. The metabolism of these oligosaccharides was also studied in growing cells metabolizing soluble sugars but also cellulose. Finally, we used liquid- and solid-state NMR to monitor wheat straw degradation by growing F. succinogenes.

Figure 1.  Pathways of formation of succinate and acetate from [1-¹³C]glucose in F. succinogenes. The position of the ¹³C-label in the ¹³C-enriched metabolite formed with [1-¹³C]glucose as substrate is shown by an asterisk. Reversions and glycogen cycling are indicated by double arrows.

Glycogen cycling and reversibility

A part of the carbohydrates metabolized by F. succinogenes may be stored as glycogen that can account for as much as 30–70% of the dry mass of the bacteria when the substrate is present in excess [7]. In strain S85, glycogen granules were seen in bacteria grown with glucose, but also cellulose (Figure 2). Glycogen was synthesized during all growth phases of the bacteria [9] and its intracellular concentration was shown to be related to the cell viability [10]. The metabolism of glycogen in this strain thus appeared to play a very important role. ¹³C and ¹H NMR were used to monitor in vivo the storage and degradation of glycogen in resting cells of F. succinogenes strain S85 [9]. The analysis of [1-¹³C]glucose metabolism in these cells showed the synthesis of glycogen labeled both on C1 (as expected) and C6 (Figures 1 and 3). The labeling at the C6 position indicates a reversion of the EMP pathway from the level of the triose phosphates (where an isomerization takes place), demonstrating the reversibility of the aldolase/triose phosphate isomerase triangle and suggesting the activity of a fructose-1,6-biphosphatase (Figure 1) [9]. We showed that about 50% of the glycogen molecules were synthesized after this reversion [11]. We also demonstrated, by using differential labeling, simultaneous storage and degradation of glycogen when bacteria were supplied with an exogenous carbon source [9]. ¹H NMR spectroscopy was used in parallel to measure the ¹³C enrichment of the end products succinate and acetate in cells metabolizing [1-¹³C]glucose, to quantify the contribution of glycogen to their synthesis, and thus estimate glycogen cycling (Figure 4) [9]. We also developed a new method, ¹³C heteronuclear spin-echo difference ¹H NMR spectroscopy, to quantify precisely the isotopic dilution [11]. This method revealed that 14% of the succinate came from unlabeled glucose, thus from endogenous glycogen. For acetate, the values obtained were much smaller than the theoretical ones, due to an additional phenomenon that was shown to be a reversal of the succinate synthesis pathway (Figure 1). This reversal was quantified, it accounted for about 10% of the acetate molecules, while 14% of the acetate came from unlabeled glucose. These results showed that 12–16% of the glucose entering the EMP pathway derived from pre-stored glycogen, and thus from glycogen cycling.

Figure 2.  F. succinogenes cells grown on glucose (A) or cellulose (B). Thin sections of cells examined by transmission electron microscopy after specific polysaccharide staining. Bars represent 0.5 µm.

Figure 3.  Spectra of in vivo kinetics of [1-¹³C]glucose utilization by resting cells of F. succinogenes S85. During the kinetics, signals of ¹³C-glucose (α and β anomers) decrease, while those of the formed metabolites (succinate, acetate, glycogen) increase. Note that glycogen is labeled on both the C1 and C6 positions.

Figure 4.  ¹H NMR spectrum of resting cells of F. succinogenes S85 metabolizing [1-¹³C]glucose. Signals due to protons attached to C2 of acetate and succinate are shown on the spectrum. Those attached to ¹²C are present in the centre of multiplets, while ¹³C-linked proton signals are split with a one bond coupling constant ¹J_13C-1H giving rise to ¹³C satellites, allowing quantification of the ¹³C/¹²C ratios.

Glycogen cycling, as well as reversal of the succinate synthesis pathway, was studied and quantified in other strains of the genus Fibrobacter[11]. The same results were obtained, showing that they were not due to a deregulation in the strain S85 resulting from extended cultivation under laboratory conditions but were a property of the genus. Such reversions and cycling were also shown in a number of other microorganisms [12]. The glycogen cycle was first described as ‘futile’ by us and other authors, but it is now clear that its role is not wasting energy. It has been suggested that the glycogen cycle is like a switch mechanism. Cells storing reserves must be able to utilize them very rapidly to adapt to sudden environmental changes. The inhibition of the ability to degrade glycogen under glycogen-accumulating conditions may prevent the efficient mobilization of the reserves when the energy demands increase abruptly. Conversely, the inhibition of glycogen synthesis under glycogen-utilizing conditions may prevent the adaptation to overflow metabolism upon restoration of favorable conditions. Therefore, cycling of storage compounds may be a key element for cells to respond quickly to environmental changes. However, the relevance of these carbohydrate cycles in vivo remains to be proven.

Synthesis of oligosaccharides

Synthesis of cellodextrins

The use of in vivo¹³C-NMR spectroscopy allowed demonstration of the polymerization of [1-¹³C]glucose into cellodextrins by resting cells of F. succinogenes. Their accumulation was important when cells were metabolizing glucose and cellobiose simultaneously [13], [14], but it was also found in cells metabolizing glucose only. As ¹H-NMR signals of saccharides are very close and often overlap, we used 2D NMR techniques (DQF COSY, HSQC, XHCORR) to analyze the oligosaccharide structure. This analysis was performed directly on extracellular medium or pelleted cells without purification. We showed that F. succinogenes resting cells metabolizing [1-¹³C]glucose produced [1-¹³C]cellodextrins that were located exclusively in the cytoplasm [15]. They had a maximum degree of polymerization (DP) of 4. These intracellular cellodextrins could be synthesized from glucose-1-phosphate generated from glycogen degradation, or from the reverse activity of a cellobiose phosphorylase, or of that of a cellobiase [15].

Synthesis of maltodextrins

We found that the bacteria metabolizing glucose were also able to synthesize and release oligosaccharides identified by 2D-NMR techniques as maltodextrins. This fact was surprising because strain S85 of F. succinogenes was not known to be able to use starch or maltose [16] and thus to have a functional maltose system. We used 2D NMR techniques and thin layer chromatography (TLC) to study the maltodextrin structure and localization. It can be stressed that the use of NMR techniques, which represents a new approach in this field, allowed simultaneous qualitative and quantitative analysis of the oligosaccharides. We found the presence of maltodextrins in both the intracellular and extracellular media. The maltodextrins were found in the millimolar concentration range in the extracellular medium, their concentration increased fourfold during the course of the incubation of resting cells with glucose. Their DP ranged from 4 to 7. The intracellular concentration of maltodextrins was estimated to be very high (up to 36 mM); it tended to decrease at the end of the incubation. The intracellular maltodextrin had a DP comprising between three and seven glucose units [15].

We also investigated the ability of resting cells of F. succinogenes S85 to use maltodextrins of different lengths. Maltose can enter the cells but is not metabolized further, as shown by TLC analysis; maltotriose is not a substrate as it is not cleaved to shorter maltose, but seems to be a key oligomer with a minimum length required for chain elongation. As the minimum DP of maltodextrins found in the extracellular medium is 4, it is suggested that maltotriose is a building block for the intracellular synthesis of maltodextrins from M4 to M6 that are then excreted [17].

Production of phosphorylated oligosaccharides

Large amounts of phosphorylated sugars including Glc-6P, Glc-1P and maltodextrin-1-phosphate (MD-1P) were identified in the 2D NMR spectra of resting cells metabolizing glucose. These compounds were present in both the intracellular and extracellular media. In the extracellular medium, the MD-1P concentration was about 0.3 mM, the DP was between 4 and 7. The intracellular concentration of MD-1P was estimated to about 4.5 mM, and the DP ranged from 3 to 7. The nature of the MD-1P was confirmed by the use of enzymes such as phosphorylase and amyloglucosidase [18].

The presence of phosphorylated sugars in the extracellular medium suggests that they are excreted by F. succinogenes.Escherichia coli is also known to be able to excrete Glc-6P under specific conditions [19]. The production of MD-1P is not common in bacteria. Intracellular maltose-1-P was shown in E. coli but its origin and role were not elucidated [20]. The phosphorylation of maltose and longer maltodextrins (DP up to 6), was also shown in Actinoplanes missouriensis[21]. In this species, MD-1P are synthesized by an ATP-dependent maltokinase that could also operate in F. succinogenes to synthesise MD-1P from MDs.

Relationship between oligosaccharide and glycogen metabolism

Analysis of the ¹²C and ¹³C ratio of cellodextrins, maltodextrins, and MD-1P showed that these oligosaccharides were synthesized from the externally provided [1-¹³C]glucose, but also from a non-labeled carbon source that could be glucose-1P or maltotriose produced from intracellular glycogen degradation [15], [18]. The precise mode of production of MD3 in F. succinogenes is not known. An enzyme with activity similar to that of MalQ (amylomaltase) of E. coli could participate in maltodextrin synthesis in F. succinogenes. However, unlike in E. coli, maltose could not be an acceptor of this 4-α-glucanotransferase if present in F. succinogenes. In E. coli mutants lacking MalQ, maltotriose is derived from the degradation of glycogen by the action of the glycogen-debranching enzyme GlgX [22], [23]. In F. succinogenes, MD3 might be produced as proposed in these E. coli mutants. A model for MD metabolism in F. succinogenes is presented in Figure 5. Like in E. coli, an ortholog of MalP (maltodextrin phosphorylase) could produce G1P from the maltodextrin.

Figure 5.  Hypothetical model for the synthesis and metabolism of maltodextrins in F. succinogenes. Maltose (M2) can enter the cells but is not metabolized further; maltotriose (M3) is not a substrate for the bacterium but is a building block for the intracellular synthesis of maltodextrins from M4 to M6 by an enzyme similar to MalQ. Maltodextrins can be excreted. Glycogen is degraded by a combination of a putative glycogen phosphorylase (GlgP, generating G-1P) and a glycogen-debranching enzyme (GlgX) that lead to maltotetraose (M4). The maltodextrin phosphorylase MalP produces M3 and G-1P from the maltodextrins.

Once the genome of F. succinogenes S85, available at TIGR (http://tigrblast.tigr.org), is completely annotated, it could be interesting to look for putative enzymes involved in intracellular MD synthesis in F. succinogenes on the basis of our model, on that of the well-known ‘maltose system’ of E. coli[24], and of the enzymes of the glycogen metabolism [25].

Modulation by the substrate

The results described above prompted us to investigate whether such maltodextrins and MD-1P were produced in bacteria metabolizing their natural substrate, cellulose. Incubations were also carried out with cellobiose, which is the main product of cellulolysis, and with a mixture of glucose and cellobiose [17]. The objective was to explore the modulation of oligosaccharide metabolism by the substrate. Cell extracts and extracellular media were analysed by 2D NMR spectroscopy and TLC. The presence of cellodextrins was looked for in these various incubations as it is assumed that cellulose degradation by rumen cellulolytic bacteria, including F. succinogenes S85, leads to the release of cellodextrins [26]. We found that no cellodextrins accumulated in extracellular media of cells whatever the substrate, and particularly cellulose. Small amounts of cellodextrins were detected inside the cells incubated with glucose, cellobiose, and also cellulose [17]. Maltodextrins (three to seven glucose units) and maltodextrin-1-phosphate were detected in incubations with cellobiose, a mixture of glucose and cellobiose, and cellulose as previously observed with incubations with glucose. A comparison of the metabolite amounts produced under the different conditions showed that, in the cells and in the extracellular medium, the phosphorylated sugar concentrations increased with the complexity of the substrate (glucose < cellobiose << cellulose). These phosphorylated sugars include: Glc-1P, MD-1P and a new metabolite named X that was only found in the case of cellulose. X was identified as a Glc-1P derivative. The increase of Glc-1P concentration in the incubations with cellobiose and cellulose may be explained by its release during the cleavage of cellobiose by the cellobiose phosphorylase [13]. The ratio MD/MD-1P decreased with the sugar size, suggesting that on cellulose the maltodextrins are more phosphorylated.

The differential accumulations of the oligosaccharides depending on the substrate address new issues about modulation of metabolism by the substrate, particularly the increase of phosphorylated species with the degree of polymerization of the β[1], [4] saccharide. This may reflect some regulation mechanism allowing adaptation of the cells to a substrate shift.

Oligosaccharide production by growing cells

Extracellular culture fluid of bacteria growing on soluble sugars and also on cellulose was analyzed by 2D NMR spectroscopy to look for the presence of oligosaccharides, phosphorylated or not, under these more physiological conditions [27]. Non-phosphorylated maltodextrins were found when cells were growing on cellulose and on the soluble sugars glucose and cellobiose. The MD-1P/maltodextrin ratio was increasing with the DP of the substrate (glucose < cellobiose << cellulose), as in resting cells. This phenomenon appeared to be amplified in growing cells where MD-1P was only transiently detected with cellulose. The Glc-1P derivative X was found under the three substrate conditions; its concentration was higher in cellulose cultures. These results, and the fact that G1P was not detected in growing cells, suggest that Glc-1P might be the precursor of X.

In F. succinogenes resting cells, no cellobiose or cellodextrins was detected in the extracellular medium of cells incubated with glucose, cellobiose or cellulose. On the contrary, cells growing on glucose accumulated cellobiose in the culture fluid, and cells growing on cellobiose accumulated cellodextrins in the culture medium [27]. These results are in accordance with those of Wells et al. [28], who showed that cellodextrins were excreted in the culture medium of cells growing on glucose or cellobiose. However, we could not detect cellobiose or cellodextrins in the medium of cellulose cultures; this suggests that cellodextrins are utilized by the cells as soon as they are released during cellulose degradation. This result is in agreement with the concept that the depolymerization of insoluble cellulose to soluble cellodextrins limits the cellulose fermentation [29].

Degradation of wheat straw by growing cells

In vitro, F. succinogenes S85 digested more cellulose from intact forages than other predominant cellulolytic bacterial species in the rumen [2]. It also has a high ability to solubilize xylans, but it is not able to use xylose [6]. The enzymatic system of F. succinogenes S85 has been extensively studied by molecular and biochemical approaches, and many different cellulases have been identified as well as xylanases, ferulic acid, acetylxylan esterase, α-arabinofuranosidase, and α-glucuronidase [2]. Although many of these enzymes have been characterized, little is known about their concurrent mode of action on solid substrates. We thus decided to study the degradation and metabolism of wheat straw by F. succinogenes cells growing on this substrate by using NMR [30]. The challenge of this work was to study both a complex substrate (natural fibers) and a complex enzymatic system (whole cells). In situ solid-state NMR, ¹³C–CP-MAS (cross-polarization magic angle spinning), was used to monitor the action of the F. succinogenes S85 fibrolytic system on lignocellulosic fibers (¹³C-enriched wheat straw), and liquid-state 2D NMR experiments were used to analyze the sugars released. A kinetic analysis of the polysaccharides degraded and of the sugars solubilized should help in defining the compounds or linkages that are limiting the degradation process by the bacterium.

The first question addressed was whether F. succinogenes cellulases degrade preferentially the amorphous regions of cellulose in wheat straw. Analysis of ¹³C–CP-MAS NMR spectra did not show a preferential degradation of amorphous regions of cellulose versus crystalline ones in wheat straw (Figure 6) [30]. This suggests either simultaneous degradation of the amorphous and crystalline parts of cellulose by the enzymes, or degradation at the surface, at the molecular scale, that cannot be detected by NMR. Indeed, it was previously shown in Clostridium cellulolyticum, a non-ruminal cellulolytic bacterium, that the crystalline and amorphous regions of cellulose were also digested at the same rate [31]. In addition, our results are in agreement with Weimer et al. [32], who suggested several years ago that the gross surface area of the fibers was a major determinant of hydrolytic rate, while substrate crystallinity was relatively unimportant.

Figure 6.  Solid-state NMR monitoring of straw degradation by F. succinogenes. ¹³C CP-MAS NMR spectra recorded after 8 h, 16 h, and 1, 2, 3, and 4 days of F. succinogenes growth. C1, C3, C4, C5, and C6 correspond to the different carbons of sugar units of cellulose and hemicelluloses. k, crystalline cellulose; nk, amorphous cellulose; OMe, CH3 of methyl ester group.

Similarly, CP-MAS NMR, and complementary chemolytic analyses showed that there was no preferential degradation of cellulose versus hemicellulose in the wheat straw (Figure 6). Again, simultaneous degradation of cellulose and hemicelluloses by the F. succinogenes enzymatic system or degradation at the surface may explain these results.

Analysis by liquid-state NMR of compounds released during F. succinogenes growth gave much more information on the mechanism of cell wall degradation by the enzymatic system. Glucose did not accumulate in the culture medium; it might not be produced from cellulose degradation, or if produced, it is used rapidly. In agreement with the fact that F. succinogenes is not able to use xylose and arabinose [6], or xylanes, hemicellulose hydrolysis products accumulated in the medium. In the rumen, they are probably rapidly used by other species that are able to metabolize them. Kinetic analysis of the production of xylose and xylo-oligosaccharides indicated very active xylanases, and identification of the nature of substituants of the xylo-oligosaccharides suggested the activity of family 10 glycosyl-hydrolases [30]. No acetylated xylan oligosaccharides could be detected, thus showing a high activity of F. succinogenes acetyl esterase. Accumulation with time of free α- and β-arabinopyranose reflected high α-arabinofuranosidase activity. Linkages between arabinose and xylose appeared to be cleaved as soon as the arabinoxylo-oligomers were produced from the xylans, indicating for the first time the activity of the F. succinogenes S85 arabinofuranosidase on a natural substrate. The high accumulation of glucuronoxylo-oligosaccharides with time suggested that the α-glucuronidase was less active than the arabinofuranosidase and acetylxylan esterase [30]. Finally, our results showed that only a low amount of 1,4-linked β-glucans was present in the culture medium after 4 days, suggesting that, as observed with cellulose cultures, cellodextrins are rapidly used by F. succinogenes S85 after their release from straw.

Conclusion

Application of various NMR techniques to the study of sugar, cellulose, and straw metabolism by the major rumen cellulolytic bacterium F. succinogenes led to important and unexpected results. We first demonstrated the occurrence of carbohydrate cycling and reversions of several metabolic reactions; such carbohydrate cycling has been proven to occur in a number of sugar-utilizing organisms, its physiological significance remains to be fully elucidated [12]. We also revealed in F. succinogenes the synthesis and excretion of maltodextrins and MD-1P and suggested their putative interconnection with glycogen metabolism. We showed that cellodextrins were not released from cellulose or straw in the extracellular medium. Several groups showed previously that F. succinogenes could be co-cultured with non-cellulolytic bacteria on cellulose, and concluded that F. succinogenes was providing cellodextrins to the other species [26], [28], [33], [34]. Indeed, it is generally admitted that cellodextrins derived from cellulose hydrolysis by cellulolytic bacteria could provide substrate to non-structural carbohydrate-fermenting cells [3], and that, within the same species, adhering cells can provide soluble carbohydrates to planktonic cells [35]. Altogether, the results obtained on growing cells indicate that maltodextrins and/or other glucose derivatives, but not cellodextrins, may be the substrate of this cross-feeding.

The origin and the role of the phosphorylated sugars accumulated by F. succinogenes cells, particularly with polysaccharide substrate, remain to be discovered. These phosphorylated sugars might well be involved in the regulation of specific gene expression, or in signaling or communication mechanisms.

In conclusion, our work illustrates the potentiality of 1D and 2D NMR techniques, applied in situ or on extracts, in liquid or solid state, to address different aspects of sugar metabolism in bacteria, including operation of metabolic cycles and metabolism of oligosaccharides and polysaccharides. We believe that case study of Fibrobacter is one of the unique examples where so many NMR approaches have been used to answer a wide variety of scientific questions. The major advantages of NMR spectroscopy are (i) that it is non-invasive and can be thus applied to complex systems and (ii) it is a method without a priori. As a consequence this allows the discovery of unexpected molecules or mechanisms. In our case it showed in particular that Fibrobacter carbohydrate metabolism is far more complex than initially thought. Integration of NMR and mass spectrometry in the study of metabolism of microorganisms, as well as genomics and proteomics, will allow microbiologists access in the near future to a complete view of the cell machinery and how it responds to changes in its environment. It would be of particular significance in the field of metabolic engineering.

Acknowledgements

We thank G. Gaudet and B. Gaillard-Martinie for the gift of the electron microscopy pictures.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.