Introduction

During growth on a mixture of sugars, bacteria will often consume the sugars sequentially through a process known as catabolite repression1,2. Usually, the cells first consume the sugar yielding the highest growth rate, followed by the sugar yielding the next highest growth rate, and so on3. They do this by repressing the expression of the genes involved in metabolizing the less preferred sugars2. This mechanism maximizes the growth rate by ensuring that cells devote their limited metabolic resources towards the preferred sugars2. The classic example involves the growth of the bacterium Escherichia coli on glucose and lactose, where E. coli will consume glucose before lactose1,4,5. The sequential utilization of these two sugars results in diauxic growth, where the cells first grow on glucose and then, following a short period of no growth, proceed to grow on lactose. This two-phase pattern of growth is also observed when E. coli is grown on mixtures of glucose and some other sugars2.

In E. coli, glucose catabolite repression is regulated by the phosphotransferase system, a multi-protein phosphorylation cascade that couples glucose uptake and metabolism4,6,7. One protein in this cascade, EIIAGlc, plays a central role in glucose catabolite repression. When EIIAGlc is phosphorylated, it activates adenylate cyclase, which increases the concentration of cAMP in the cells. cAMP activates the expression of many genes involved in the metabolism of non-glucose sugars through the action of the cAMP receptor protein CRP. In addition, the expression of these genes is also induced by their cognate sugars. Because EIIAGlc phosphorylation increases when the cells are not metabolizing glucose, these CRP-regulated metabolic genes are expressed only in the absence of glucose and presence of their cognate sugar. In addition, unphosphorylated EIIAGlc can bind and inhibit the ability of some sugar transporters to uptake their cognate sugars through a process known as inducer exclusion4,6,7,8,9,10. Thus, when the cells are metabolizing glucose and EIIAGlc is predominantly in the unphosphorylated form, these sugars are not transported into the cell and, therefore, are unable to induce the expression of their cognate metabolic genes. These two mechanisms are believed to be the main factors regulating glucose catabolite repression in E. coli.

Catabolite repression has also been observed with sugar mixtures not involving glucose. One prominent example is the growth of E. coli on mixtures of L-arabinose and D-xylose (hereafter referred to simply as arabinose and xylose)11. In this case, the cells will consume arabinose before xylose, though the extent of repression is less pronounced than observed with glucose and arabinose or glucose and xylose. Arabinose catabolite repression results from regulatory crosstalk between these two sugar utilization systems12. The arabinose transport and metabolic genes are regulated by the transcription factors AraC and CRP. When AraC is bound with arabinose and CRP with cAMP, they activate the expression of the arabinose genes. Similarly, the xylose transport and metabolic genes are regulated by the transcription factors XylR and CRP. When XylR is bound with xylose and CRP with cAMP, they activate the expression of the xylose transport and metabolic genes. Arabinose catabolite repression results from arabinose-bound AraC binding to the promoters of the xylose genes and inhibiting their expression through a competitive mechanism. In addition, xylose weakly represses the expression of the arabinose genes using an analogous mechanism involving xylose-bound XylR. This reciprocal mechanism generates a small subpopulation of the cells that selectively consume xylose even in the presence of arabinose13.

A recent study by Aidelberg and coworkers investigated the utilization of six non-glucose sugars (lactose, arabinose, xylose, D-sorbitol, L-rhamnose and D-ribose) in E. coli14. They observed hierarchical expression of the metabolic genes for these sugars. In particular, they found that sugars higher up in the hierarchy repress the expression of the genes involved in the metabolism of the sugars lower down in the hierarchy. These results suggest that the cells selectively utilize these six sugars in a similar hierarchical manner. In addition, they proposed that the hierarchical expression of these genes is regulated by CRP through the differential production of cAMP. A number of studies have shown that cAMP levels are inversely proportional to the growth rate of the cells, irrespective of the sugar, due to inhibition of adenylate cyclase by the downstream metabolites15,16,17. In support of this hypothesis, Aidelberg and coworkers found that the expression hierarchy matches the growth rates associated with these sugars. Furthermore, they found that the promoters for genes associated with sugars at the top of the hierarchy are more sensitive to cAMP than those at the bottom of the hierarchy. In other words, when the cell is growing on a higher-growth yielding sugar such as lactose, insufficient cAMP is produced to induce the expression of genes for lower-growth yielding sugars such as rhamnose. This mechanism is appealing as it readily explains how cells selectively utilize sugars in accords with their growth yield. These results also suggest that multiple mechanisms are involved in the selective utilization of non-glucose sugars, where some involve transcriptional crosstalk, as is the case with arabinose and xylose, and others do not.

In this work, we investigated the selective utilization of lactose, arabinose, and xylose. When E. coli is grown on a mixture of lactose and arabinose or lactose and xylose, it will consume the lactose first. Motivated by the study of Aidelberg and coworkers, we sought to determine whether the preferential utilization of lactose is due to cAMP or regulatory crosstalk. Using a number of different approaches, we conclude that the selective utilization of lactose is mainly due to cAMP. These results indicate that E. coli employs multiple mechanisms to regulate the utilization of non-glucose sugars.

Results

Lactose represses arabinose and xylose gene expression

As a first step towards investigating catabolite repression in sugar mixtures involving lactose and arabinose or lactose and xylose, we measured how these sugars individually affect the expression of each other’s metabolic genes. To measure gene expression, we employed single-copy transcriptional fusions to the fluorescent protein Venus. Briefly, we grew the cells in TB7 medium containing 1 mM of the cognate sugar, which is sufficient to fully activate the expression of the corresponding metabolic genes, and increasing concentrations of a second sugar. As shown in Fig. 1a, lactose inhibited expression from both the araB and xylA promoters, which respectively control the expression of the arabinose and xylose metabolic genes. These results are consistent with the results reported by Aidelberg and coworkers14. In addition, we found that lactose represses the xylA promoter more strongly than the araB promoter.

Figure 1
figure 1

Reciprocal repression of lactose, arabinose, and xylose gene expression. (a) Effect of increasing concentrations of lactose on araB (grey bars) and xylA (white bars) promoter activity as determined using transcriptional fusions to the fluorescent protein Venus (P araB -Venus and P xylA -Venus). Cells were grown in TB7 with 1 mM arabinose or xylose. (b) Effect of increasing arabinose concentrations on lacZ promoter activity as determined using transcriptional fusions to the fluorescent protein Venus (P lacZ -Venus). Cells were grown in TB7 with 1 mM lactose. (c) Effect of increasing xylose concentrations on lacZ promoter activity as determined using transcriptional fusions to the fluorescent protein Venus (P lacZ -Venus). Cells were grown in TB7 with 1 mM lactose. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image

We also tested whether arabinose and xylose inhibit the lacZ promoter, which controls the expression of the lactose metabolic genes. As shown in Fig. 1b, arabinose represses expression from the lacZ promoter. These results indicate that the repression is reciprocal. In other words, lactose represses arabinose metabolism and arabinose represses lactose metabolism. However, repression by lactose is somewhat greater than repression by arabinose. In the case of xylose (Fig. 1c), we found that it weakly represses expression from the lacZ promoter. While repression is also reciprocal during growth on lactose and xylose, lactose is clearly the preferred sugar.

We also performed identical experiments in M9 minimal medium13,18. Similar to what was observed in TB7, we found that lactose inhibits the expression of the arabinose and xylose metabolic genes in M9 minimal medium (Figure S1), though the effect is stronger in this medium. In addition, we found that arabinose again weakly represses the expression of the lactose metabolic genes. However, xylose was found to have no significant effect on the expression of the lactose genes in M9 minimal medium.

Lactose reduces the rate of arabinose and xylose consumption

We next investigated sugar utilization during growth on arabinose or xylose in the presence of lactose. We grew the cells in sugar mixtures containing equal concentrations on a molar basis of the two sugars (2 mM). As shown in Fig. 2, the cells individually consume lactose and arabinose at roughly the same rates. However, the rate of utilization for both sugars was decreased somewhat when the cells were grown in the presence of both. These results are consistent with a mechanism of reciprocal repression. Only during growth on lactose and xylose was catabolite repression observed. In this case, lactose significantly reduced the rate of xylose consumption. However, xylose did not reduce lactose consumption. Unlike arabinose, xylose is intrinsically consumed more slowly than lactose. Moreover, the rate of xylose consumption is further reduced in the presence of lactose, most likely due to repression of the xylose metabolic genes. The results are consistent with our gene expression results (Fig. 1), where lactose exhibits a greater effect on xylose genes than on arabinose genes.

Figure 2
figure 2

Lactose inhibits the utilization of arabinose and xylose. (a) Sugar utilization in mixtures of lactose and arabinose. Diamonds refer to arabinose concentrations in the presence (black) or absence (white) of lactose; circles refer to lactose concentrations in the presence (black) or absence (white) of arabinose. (b) Sugar utilization in mixtures of lactose and xylose. Diamonds refer to xylose concentrations in the presence (black) or absence (white) of lactose; circles refer to lactose concentrations in the presence (black) or absence (white) of xylose. Cells were grown in TB7. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image

We also measured the sugar utilization in M9 minimal medium. As shown in Figure S2, lactose delayed the consumption of both arabinose and xylose. These results are consistent with those obtained using TB7 (Fig. 2). However, in M9 minimal medium, arabinose did not delay the utilization of lactose, unlike the case with TB7 medium. In addition, both lactose and xylose were utilized at similar rates when they were individually added to the growth medium. These results demonstrate that lactose represses both the utilization of arabinose and xylose in M9 minimal medium. However, reciprocal repression does not occur in this medium.

Lactose catabolite repression does not involve transcriptional crosstalk

Previous work demonstrated that arabinose represses xylose metabolism and vice versa through a mechanism involving transcriptional crosstalk12. In particular, arabinose-bound AraC binds the xylA promoter and represses transcription. Similarly, xylose-bound XylR binds the araB promoter and represses transcription, albeit to a lesser extent. One possibility is that lactose represses arabinose and xylose gene expression also using competitive mechanism. Briefly, the transcriptional repressor LacI regulates the expression of the lactose genes. When LacI is bound with allolactose, a by-product of lactose metabolism, it no longer binds the lacZ promoter and represses transcription. We therefore hypothesized that allolactose-bound LacI binds to the araB and xylA promoters and represses transcription of the respective metabolic genes. To test this hypothesis, we measured expression from the araB and xylA promoters in a ΔlacI mutant. As shown in Fig. 3, lactose was still able to repress araB and xylA promoters in the absence of LacI. These results demonstrate that lactose catabolite repression does not involve transcriptional crosstalk by LacI.

Figure 3
figure 3

Lactose catabolite repression does not involve transcriptional crosstalk by LacI. (a) Effect of increasing concentrations of lactose on araB promoter activity in ΔlacI mutant (ΔlacI P araB -Venus). Cells were grown in TB7 with 1 mM arabinose. (b) Effect of increasing concentrations of lactose on xylA promoter activity in ΔlacI mutant (ΔlacI P xylA -Venus). Cells were grown in TB7 with 1 mM xylose. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image

Lactose-mediated repression is independent of arabinose and xylose metabolism

We next tested whether lactose is able to repress arabinose and xylose gene expression in the absence of arabinose and xylose metabolism. The goal here was to see whether lactose was affecting some downstream metabolic process that somewhat would affect arabinose and xylose gene expression. As shown in Figs 4a and 5a, lactose still inhibited arabinose and xylose gene expression in strains unable to metabolize these respective sugars. These results demonstrate that lactose-mediated repression of the arabinose and xylose genes is independent of arabinose and xylose metabolism.

Figure 4
figure 4

Lactose-mediated repression is independent of arabinose metabolism. (a) Effect of increasing concentrations of lactose on araB promoter activity in ΔaraBAD mutant (ΔaraBAD P araB -Venus). Cells were grown in TB7 with 1 mM arabinose. (b) Effect of increasing concentrations of arabinose on lacZ promoter activity in ΔaraBAD mutant (ΔaraBAD P lacZ -Venus). Cells were grown in TB7 with 1 mM lactose. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image
Figure 5
figure 5

Lactose-mediated repression is independent of xylose metabolism. (a) Effect of increasing concentrations of lactose on xylA promoter activity in ΔxylAB mutant (ΔxylAB P xylA -Venus). Cells were grown in TB7 with 1 mM xylose. (b) Effect of increasing concentrations of xylose on lacZ promoter activity in ΔxylAB mutant (ΔxylAB P lacZ -Venus). Cells were grown in TB7 with 1 mM lactose. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image

Arabinose and xylose-mediated repression of the lactose metabolism genes is respectively dependent on arabinose and xylose metabolism

We next tested whether the repression of the lactose genes by arabinose and xylose was respectively dependent on arabinose and xylose metabolism. As shown in Figs 4b and 5b, we observed no repression of lactose gene expression in the absence of arabinose and xylose metabolism. The results demonstrate that arabinose or xylose metabolism is necessary for the repression of the lactose genes. They also suggest that the mechanism for repression involves cAMP. As the respective regulators, AraC and XylR, were still present in these mutants, these results also indicate that the mechanism does not involve transcriptional crosstalk.

We also performed the reciprocal experiment, where we tested whether lactose is able to repress arabinose and xylose gene expression in the absence of lactose metabolism using ΔlacZ mutant. Not surprisingly, we observed no repression (data not shown). However, these experiments were uninformative, because they provide no insight regarding the mechanism of lactose-mediated repression. The reason is that LacZ, the enzyme that cleaves lactose into glucose and galactose, also produces allolactose, the inducer for LacI19. Thus, we cannot determine whether loss of repression in a ΔlacZ mutant results from the loss of the LacI inducer or the inability of the cells to utilize lactose.

Repression by lactose can be explained by preferential promoter activation by cAMP-CRP

Aidelberg and coworkers previously proposed that the hierarchy among lactose, arabinose and xylose is due to differential promoter activation by the CRP global regulator14. To further test this hypothesis, we measured whether the addition of cAMP would relieve lactose inhibition of the arabinose and xylose promoters. As shown in Fig. 6, the addition of external cAMP increased araB and xylA promoter activities in presence of lactose. In the case of the araB promoter, repression was relieved by the addition of 5 or 10 mM cAMP. However, in the case of the xylA promoter, repression was only slightly relieved, even though similar concentrations of cAMP were employed. We were unable to test higher concentrations of cAMP (>10 mM), because they hindered cell growth. Nonetheless, these results are consistent with a mechanism whereby lactose inhibits arabinose and xylose metabolism through CRP.

Figure 6
figure 6

cAMP mitigates lactose-mediated repression of arabinose and xylose gene expression. (a) Effect of different concentrations of cAMP on araB promoter activity. Cells were grown in M9 minimal medium containing 1 mM arabinose. (b) Effect of different concentrations of cAMP on xylA promoter activity. Cells were grown in M9 minimal medium containing 1 mM xylose. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image

We also tested whether the addition of cAMP would relieve the repression of the lacZ promoter by arabinose. As shown in Figure S3, we found that repression was partially relieved by the addition of 5 mM cAMP. Higher concentrations of cAMP (10 mM), however, did not further increase the activity of the lacZ promoter. These results demonstrate that arabinose-mediated repression of the lacZ promoter is due in part to cAMP. Whether it represents the sole mechanism is unknown.

Swapping the CRP binding site of xylA promoter with that of lacZ promoter reduces lactose-mediated repression of xylose metabolism

Previous results suggest that the CRP binding site within the lacZ promoter has a higher affinity for cAMP-CRP than the CRP binding site within the xylA promoter14. If catabolite repression is due to differential activation by CRP, then replacing the weak CRP binding site within the xylA promoter with a stronger one from the lacZ promoter should relieve repression. The xylA promoter has a single CRP binding site20. The lacZ promoter has two CRP binding sites21,22. The distal site is considered to be the main CRP binding site for the lacZ promoter. Therefore, we replaced the native CRP binding site within the xylA promoter with the distal CRP binding site from the lacZ promoter. As shown in Fig. 7, less repression by lactose was observed with the xylA promoter containing the distal lacZ CRP binding site than with the one containing the native CRP binding site. These results further support a mechanism where lactose-mediated repression of xylose metabolism is due to cAMP.

Figure 7
figure 7

Swapping the CRP binding site of xylA promoter with that of lacZ promoter reduces lactose-mediated repression of xylose metabolism. Effect of increasing concentrations of lactose on the activity of xylA promoter with native (grey bars) or lacZ (white bars) CRP binding site. Cells were grown on M9 minimal medium with 1 mM xylose. Error bars denote the standard deviation of three experiments performed on separated days.

Full size image

Discussion

E. coli is capable of growing on a number of different sugars. While much is known about how E. coli regulates its metabolism when grown on a single sugar, far less is known about how this bacterium regulates its metabolism when grown on mixtures of sugars, particularly when the mixture does not involve glucose. A number of studies have shown that E. coli will selectively utilize non-glucose sugars, first consuming the sugar that yields the highest growth rate and then consuming the sugar that yields the next highest growth and so on11,12,14. Still many questions remain about the mechanisms governing the selective utilization of these sugars. Two mechanisms have recently been proposed. One involves transcriptional crosstalk between the regulatory systems governing the metabolism of individual sugars. This mechanism was shown to govern the preferential utilization of arabinose over xylose12. The other mechanism involves cAMP, a second messenger involved in glucose catabolite repression4,6. This mechanism was proposed to govern the preferential utilization of multiple non-glucose sugars14. However, this study did not consider whether transcriptional crosstalk was also involved. Indeed, cAMP does not appear to govern the preferential utilization of arabinose over xylose12. Furthermore, repression is reciprocal in the case of arabinose and xylose, where arabinose strongly represses the expression of the xylose metabolic genes and xylose weakly represses the expression of the arabinose metabolic genes12. These results indicate that the hierarchy among sugars is not necessarily strict but rather depends on the relative concentrations of the two sugars.

In this work, we investigated the selective utilization of three sugars in E. coli: lactose, arabinose, and xylose. Our goal was to determine whether transcriptional crosstalk or cAMP/CRP governed the selective utilization of these sugars. In rich medium, we found that both lactose and arabinose inhibited the utilization of the other (Fig. 2b). Consistent with these results, we found that lactose inhibits the expression of the arabinose metabolic genes and that arabinose inhibits the expression of the lactose metabolic genes. In minimal medium, however, lactose is the preferred sugar. These results are also consistent with gene expression data (Figure S1), where lactose is a stronger repressor of arabinose gene expression than arabinose is of lactose gene expression. The origin of these differences on rich and minimal media is not known, though may reflect the complex growth patterns observed in peptide-based medias23,24. In the case of xylose and lactose, E. coli prefers the latter in both complex and minimal media. Even though xylose inhibits the expression of the lactose genes, the effect is relatively weak.

This hierarchy matches the growth rates associated with these sugars14. One question is why does E. coli grow more rapidly on lactose than on arabinose or xylose. The increased growth rate associated with lactose is likely due to its metabolism requiring less steps: intracellular lactose is degraded by the enzyme LacZ into galactose and glucose, where the latter can directly enter the Embden–Meyerhof–Parnas (glycolytic) pathway. Arabinose and xylose metabolism involves many more steps, as they both enter the glycolytic pathway through the non-oxidative branch of the pentose phosphate pathway. This likely explains why their metabolism is more inefficient, in addition to their lower energetic yield. Therefore, one would expect that lactose inhibits the metabolism of arabinose and xylose, as this would reflect a more efficient allocation of limited metabolic resources.

Regarding the mechanism, our data support one involving cAMP. Indeed, we found that lactose was still able to repress the expression of the arabinose and xylose metabolic genes when LacI, the cognate regulator of the lactose genes25, was deleted (Fig. 3). Had the mechanism involved transcriptional crosstalk, we would expect repression to be eliminated in a ΔlacI mutant, contrary to what was observed. These results indicate that lactose metabolism alone is sufficient for repressing the expression of the arabinose and xylose genes. Likewise, we found that arabinose and xylose metabolism were necessary for arabinose and xylose to repress lactose gene expression (Figs 4 and 5).

That metabolism is necessary for repression strongly suggests that cAMP is involved. In support of this mechanism, we found that the addition of exogenous cAMP partially relieved lactose-mediated repression of arabinose and xylose gene expression (Fig. 6). In the case of the arabinose genes, exogenous cAMP was nearly able to counteract the repressive effects of lactose. In the case of the xylose genes, however, exogenous cAMP was only able to partially relieve the repression by lactose, possibly due to our inability to explore higher cAMP concentrations15,16,17. As a further test, we also replaced the native CRP binding site within the xylA promoter with one from the lacZ promoter. Consistent with a mechanism involving cAMP, we found that lactose-mediated repression was reduced when the CRP site was replaced in the xylA promoter (Fig. 7).

While our results demonstrate that cAMP regulates the sugar utilization in mixtures of lactose and arabinose or xylose, it may not be the only mechanism involved. One possibility is that allolactose, a byproduct of lactose metabolism and the inducer for LacI19, could inhibit AraC and XylR, the respective regulators of arabinose and xylose gene expression20,26,27. Previous work has shown that isopropyl β-D-1-thiogalactopryanoside (IPTG), a non-metabolizable lactose analog, competitively inhibits AraC28. We tested whether it competitively inhibits XylR. However, we found that it had no effect on xylose gene expression (data not shown). We also tested whether thiomethyl galactoside (TMG), another non-metabolizable lactose analog, inhibits arabinose and xylose gene expression (Figure S4). Consistent with the IPTG results, we found that TMG competitively inhibits arabinose gene expression but not xylose gene expression (Figure S4). These results suggest that allolactose may competitively inhibit AraC, though whether these non-metabolizable analogs truly mimic the effect of allolactose is unknown. Unfortunately, we were unable to directly test allolactose as this sugar is difficult to obtain. Nevertheless, these results indicate that allolactose likely does not repress XylR. Furthermore, as lactose represses xylose gene expression more strongly than arabinose gene expression, it seems unlikely that inhibition of AraC by allolactose, assuming it occurs, is a significant contributor to catabolite repression.

Why does E. coli employ different mechanisms for selective sugar utilization? As noted above, the mechanism differs for a mixture involving arabinose and xylose and those involving lactose and arabinose or lactose and xylose. One possibility is that these differences reflect the relative frequency of encountering these different sugar mixtures in nature. In particular, E. coli is unlikely to encounter sugar mixtures involving lactose and arabinose or lactose and xylose. Lactose is found in milk whereas arabinose and xylose are found in the complex polysaccharides that comprise the plant cell wall29. In other words, cells are unlikely to be exposed to mixture of lactose and arabinose or lactose and xylose whereas they will likely be exposed to mixture of arabinose and xylose. This would suggest that cAMP is a general mechanism for catabolite repression whereas transcriptional crosstalk is a more specialized one. In particular, mechanisms involving transcriptional crosstalk likely evolved to regulate more frequent events than the generalized mechanism involving cAMP, which appears not to be adapted for any specific sugars.

In summary, we investigated the mechanism for catabolite repression in sugar mixtures of lactose, arabinose, and xylose. Building on previous work, we established that cAMP principally regulates the selective utilization of lactose and arabinose; and lactose and xylose. Moreover, we found that repression is reciprocal in the sense that both sugars repress each other’s metabolism, though lactose is clearly the dominant sugar. These results further our understanding of metabolism in E. coli. They also demonstrate that multiple mechanisms are involved in regulating the metabolism of sugar mixtures not involving glucose.

Methods

Media and growth conditions

Luria-Bertani (LB) medium (10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl) was used for strain construction. All experiments were performed in either buffered tryptone broth (TB7: 10 g/L tryptone and 1 mM MgSO4, buffered at pH 7 with 100 mM potassium phosphate) or M9 minimal medium (6.8 g/L Na2HPO4, 3 g/L KH2PO4, 1 g/L NH4Cl, 0.5 g/L NaCl, 2 mM MgSO4, 100 μM CaCl2, 0.001% thiamine hydrochloride and 4 g/L glycerol). All experiments were performed at 37 °C. Antibiotics were added as needed at the following concentrations: ampicillin, 100 μg/mL; chloramphenicol, 20 μg/mL; kanamycin, 40 μg/mL.

Strain and plasmid construction

Table 1 lists all strains and plasmids used in this study. Table S1 lists all oligonucleotides used in strain and plasmid construction. All strains are isogenic derivatives of E. coli MG1655. Genes were deleted using the method of Datsenko and Wanner30. Prior to removing the antibiotic resistant cassette using the FLP recombinase, P1 transduction was used to move the cassette to a clean parental background. The plasmid pXW211 contains a transcriptional fusion of the xylA promoter to the fluorescent protein Venus. To construct this plasmid, xylA promoter (genomic region: 3730691–3731015) was PCR amplified using the primers SK435F and XW435R, and then the PCR fragment was cloned into the integrative plasmid pVenus using SalI and EcoRI restriction sites. The plasmid pSK460 contains a transcriptional fusion of the lacZ promoter to the fluorescent protein Venus. To construct this plasmid, the lacZ promoter (genomic region: 366222–366545) was PCR amplified using the primers SK418F and XW418R, and then the PCR fragment was cloned into the integrative plasmid pVenus using SalI and EcoRI restriction sites. The plasmid pXW238 contains a transcriptional fusion of the xylA promoter to the fluorescent protein Venus, where the native CRP binding site of the xylA promoter (TTTTGCGAGCGAGCGCACACTT) (genomic region: 3730899–3730878)20,31 was replaced with the CRP binding site of the lacZ promoter (TAATGTGAGTTAGCTCACTCAT) (genomic region: 366394–366415)32. To construct this plasmid, the plasmid pXW211 was PCR amplified using the primers EA061F and EA062R, which contain the CRP binding site from the lacZ promoter as 5′ overhangs. The resulting PCR fragment was then phosphorylated using T4 polynucleotide kinase and ligated back into a circular plasmid using T4 DNA ligase. The plasmid was then integrated into λ phage attachment sites using the CRIM integration system33. After confirming by PCR that a single integration took place, the integrated plasmid was moved into a clean background using P1 transduction34,35.

Table 1 Strains and plasmids used in this study.
Full size table

Fluorescence assays

In the experiments involving TB7 as the base medium, the cultures were first grown overnight in TB7. The overnight culture was then diluted 1:100 into fresh media and grown for 2 h to an OD600 of ~0.2. The cultures were then subcultured into 1 mL of fresh TB7 medium containing the specified sugar to an initial OD600 of 0.01 in 96 deep-well plates. The cultures were then incubated on a microplate shaker at 600 rpm for 5 h. In the experiments involving M9 as the base media, the cells were grown overnight in M9 minimal media. The overnight culture was diluted 1:50 into fresh media and grown for 5 h to an OD600 of ~0.2. One mL aliquots were then distributed to individual wells in a 96 deep-well plate. The specified sugars were then added to each well, and then the plate was incubated on the microplate shaker at 600 rpm for 4 h. At the end of every experiment, 100 µL samples were taken from each well and transferred to a black, clear-bottom Costar 96-well microplate. Fluorescence (excitation at 515 nm; emission at 528 nm) and absorbance (OD600) were measured using Tecan Safire2 microplate reader. The fluorescence readings were normalized with the OD600 to account for cell density. All experiments were performed on three separate days, and the average values with the standard deviations are reported.

Analytical measurements

Cells were first grown overnight in the specified media. The overnight culture was then sub-cultured into 60 mL of fresh media containing the specified sugars in a 125 mL shake flask with an initial OD600 of 0.05. The flasks were then incubated on a standard orbital shaker at 250 rpm and 37 °C. Samples (1 mL) were taken at specified time intervals and filtered through 0.22 μm filters (Millipore) for subsequent analysis. Arabinose, lactose and xylose concentrations were measured using a Shimadzu HPLC system equipped with Shimadzu RID-10A refractive index detector (Shimadzu Corp., Kyoto, Japan), and an Aminex HPX-87H column (300 mm × 7.8 mm) (Bio-Rad Laboratories, Hercules, CA, USA). The HPLC oven was adjusted to 65 °C and a mobile phase of 0.5 mM H2SO4 was run at 0.6 mL/min for 15 min per sample. Kinetic experiments were run in duplicates with the averages and standard deviations reported.

Data availability

Any other datasets generated during and/or analyzed during the current study are available from the corresponding author upon request.