bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
- 1 The impact of CO2/HCO3 availability on anaplerotic flux in PDHC-
2 deficient Corynebacterium glutamicum strains
3
a a a a b# 4 Aileen Krüger , Johanna Wiechert , Cornelia Gätgens , Tino Polen , Regina Mahr
5 and Julia Frunzkea#
6
7 a Institut für Bio- und Geowissenschaften, IBG-1: Biotechnology, Forschungszentrum
8 Jülich, 52425 Jülich, Germany
9 b SenseUp GmbH, c/o Campus Forschungszentrum Jülich, 52425 Jülich, Germany
10
- 11 Running Head: Influence of CO2/HCO3 on anaplerotic flux
12
13 A.K. and J.W. contributed equally to this work.
14
15 #Address correspondence to Julia Frunzke, [email protected], and Regina
16 Mahr, [email protected].
17
1
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
18 Abstract
19 The pyruvate dehydrogenase complex (PDHC) catalyzes the oxidative
20 decarboxylation of pyruvate yielding acetyl-CoA and CO2. The PDHC-deficient
21 Corynebacterium glutamicum strain ΔaceE is therefore lacking an important
22 decarboxylation step in central metabolism. Additional inactivation of pyc, encoding
23 pyruvate carboxylase, resulted in a >15 hour lag phase in the presence of glucose,
24 while no growth defect was observed on gluconeogenetic substrates like acetate.
25 Growth was successfully restored by deletion of ptsG encoding the glucose-specific
26 permease of the PTS system, thereby linking the observed phenotype to the
27 increased sensitivity of strain ΔaceE Δpyc to glucose catabolism. In the following,
28 strain ΔaceE Δpyc was used to systematically study the impact of perturbations of
- 29 the intracellular CO2/HCO3 pool on growth and anaplerotic flux. Remarkably, all
- - 30 measures leading to enhanced CO2/HCO3 levels, such as external addition of HCO3
31 , increasing the pH, or rerouting metabolic flux via pentose phosphate pathway, at
32 least partially eliminated the lag phase of strain ΔaceE Δpyc on glucose medium. In
- 33 accordance, inactivation of the urease enzyme, lowering the intracellular CO2/HCO3
34 pool, led to an even longer lag phase accompanied with the excretion of L-valine and
35 L-alanine. Transcriptome analysis as well as an adaptive laboratory evolution
36 experiment of strain ΔaceE Δpyc revealed the reduction of glucose uptake as a key
37 adaptive measure to enhance growth on glucose/acetate mixtures. Altogether, our
- 38 results highlight the significant impact of the intracellular CO2/HCO3 pool on
39 metabolic flux distribution, which becomes especially evident in engineered strains
40 suffering from low endogenous CO2 production rates as exemplified by PDHC-
41 deficient strains.
42
2
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
43 Importance: CO2 is a ubiquitous product of cellular metabolism and an essential
44 substrate for carboxylation reactions. The pyruvate dehydrogenase complex (PDHC)
- 45 catalyzes a central metabolic reaction contributing to the intracellular CO2/HCO3 pool
46 in many organisms. In this study, we used a PDHC-deficient strain of
47 Corynebacterium glutamicum, which was additionally lacking pyruvate carboxylase
48 (ΔaceE Δpyc). This strain featured a >15 h lag phase during growth on glucose-
49 acetate mixtures. We used this strain to systematically assess the impact of
- 50 alterations in the intracellular CO2/HCO3 pool on growth on glucose-containing
- 51 medium. Remarkably, all measures enhancing the CO2/HCO3 levels successfully
- 52 restored growth emphasizing the strong impact of the intracellular CO2/HCO3 pool on
53 metabolic flux especially in strains suffering from low endogenous CO2 production
54 rates.
55
56
3
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
57 Introduction
58 CO2 is an inevitable product and at the same time an essential substrate of microbial
- 2- 59 metabolism. In water, CO2 is in equilibrium with HCO3 and CO3 , which is influenced
60 by the pH of the medium (1). As a product of decarboxylating reactions and substrate
- 61 for carboxylations, CO2 and bicarbonate (HCO3 ) are involved in various metabolic
- 62 processes. Consequently, the intracellular CO2/HCO3 pool has a significant impact
63 on metabolic fluxes like e.g. anaplerosis, especially in the early phase of cultivations
64 when cell density is low (2).
65 During aerobic growth, the pyruvate dehydrogenase complex (PDHC), which is
66 conserved in various microbial species, catalyzes an important reaction contributing
- 67 to the intracellular CO2/HCO3 pool (3, 4). The PDHC is a multienzyme complex
68 belonging to the family of 2-oxo acid dehydrogenase complexes also comprising the
69 2-ketoglutarate dehydrogenase complex (ODHC) and the branched-chain 2-ketoacid
70 dehydrogenase (3, 5). In particular, PDHC catalyzes the oxidative decarboxylation of
71 pyruvate to acetyl-CoA, thereby producing one molecule CO2 and one molecule
72 NADH per mol pyruvate.
73 The Gram-positive actinobacterium Corynebacterium glutamicum represents an
74 important platform strain used in industrial biotechnology for the production of amino
75 acids, proteins and various other value-added products (6-10). In this organism, the
76 PDHC complex represents a central target to engineer metabolic pathways of
77 pyruvate-derived products including L-valine, isobutanol, ketoisovalerate and L-
78 lysine. To this end, different studies focused on the reduction or complete
79 abolishment of PDHC activity to improve precursor availability (4, 11-13). Due to the
80 deficiency of PDHC activity, however, cells are not able to grow on glucose as single
81 carbon source which can be circumvented by the addition of acetyl-CoA refueling
82 substrates such as acetate. In presence of both carbon sources glucose and acetate, 4
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
83 PDHC-deficient strains initially form biomass from acetate and subsequently, convert
84 glucose into products (e.g. L-valine or L-alanine) in the stationary phase (12). In
85 contrast to Escherichia coli or Bacillus subtilis, C. glutamicum does not normally
86 show the typical diauxic growth behaviour, but prefers co-metabolization of many
87 carbon sources (10, 14). Co-utilization of glucose and acetate has been studied in
88 detail, showing that consumption rates of both carbon sources decrease, but the total
89 carbon consumption is comparable to growth on either carbon source alone which is
90 regulated by SugR activity (15). Further examples for co-metabolization of glucose
91 with other sugars or organic acids include fructose (16), lactate (17) as well as
92 pyruvate (14) and gluconate (18). It was shown, that upon growth on glucose-acetate
93 mixtures, the glyoxylate shunt is required to fuel the oxaloacetate pool in the wild type
94 (15). Here it is important to note that the glyoxylate shunt is active in the presence of
95 acetate, but repressed by glucose (19, 20).
96 For growth on glycolytic carbon sources, organisms depend on TCA cycle
97 replenishing reactions constituting the anaplerotic node (or phosphoenolpyruvate-
98 pyruvate-oxaloacetate node), comprised of different carboxylating and
99 decarboxylating reactions (21). Consequently, flux via these reactions is influenced
- 100 by the intracellular CO2/HCO3 pool. In contrast to most other organisms, C.
101 glutamicum possesses both anaplerotic carboxylases namely the
102 phosphoenolpyruvate carboxylase (PEPCx) (EC 4.1.1.31) encoded by the ppc gene
103 (22, 23) and the pyruvate carboxylase (PCx) (EC 6.4.1.1) encoded by pyc (21, 24).
104 These two C3-carboxylating enzymes catalyze bicarbonate-dependent reactions
105 yielding oxaloacetate from PEP or pyruvate, respectively. In C. glutamicum, it was
106 shown that these two enzymes could replace each other to a certain extent
107 depending on the intracellular concentrations of the respective effectors for each
108 enzyme. However, under standard conditions during growth on glucose, the biotin- 5
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
109 containing PCx contributes to the main anaplerotic activity of 90% compared to
110 PEPCx (24). Remarkably, the Michaelis-Menten constants of both carboxylases are
111 about 30-fold higher in comparison to Escherichia coli PEP carboxylase (25).
- 112 Apparently, a low CO2/HCO3 pool may thus limit anaplerotic flux, which is of special
113 relevance in early phases when biomass concentration is low, but high aeration may
114 strip dissolved CO2 from the medium.
115 Several studies revealed the inhibitory effect of low pCO2 on microbial growth (2, 26,
116 27). In the case of C. glutamicum, low pCO2 pressures led to a significant drop in
117 growth rate in turbidostatic continuous cultures (28) as well as batch cultures (29). It
118 was further demonstrated that the reduced flux through anaplerotic reactions under
- 119 low CO2/HCO3 levels led to an increased production of the pyruvate-derived amino
120 acids L-alanine and L-valine (29).
121 While several previous studies focused on the impact of CO2 by altering the CO2
122 partial pressure in the process, we applied targeted genetic perturbations to
- 123 systematically assess the impact of the intracellular CO2/HCO3 pool on anaplerotic
124 flux and growth of C. glutamicum. Here, we focused on PDHC-deficient strains
- 125 suffering from a reduced intracellular CO2/HCO3 pool during growth on glucose-
- 126 acetate mixtures. The effect of low CO2/HCO3 levels became even more evident,
127 upon additional deletion of the pyc gene encoding the dominant anaplerotic enzyme
128 leading to a drastically elongated lag phase of a C. glutamicum ΔaceE Δpyc strain
129 during growth on glucose acetate mixtures. This effect could be attributed to a
130 reduced tolerance of these strains towards glucose and was successfully
131 complemented by deletion of ptsG encoding the glucose-specific permease of the
132 PTS system. Remarkably, growth was successfully restored by increasing the
- 133 intracellular CO2/HCO3 pool by the addition of bicarbonate, by increasing the pH, by
134 re-routing metabolic flux over the pentose phosphate pathway or by refueling of the 6
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
135 oxaloacetate pool by the addition of TCA intermediates to the growth medium.
136 Finally, an adaptive laboratory evolution (ALE) of C. glutamicum ΔaceE Δpyc strain
137 on minimal medium containing glucose and acetate revealed the abolishment of
138 glucose uptake as a primary strategy to allow for growth on acetate.
139
140
7
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
141 Results
- 142 A lowered intracellular CO2/HCO3 pool impacts growth of PDHC-deficient
143 strains
144 The pyruvate dehydrogenase complex (PDHC) comprises a key metabolic reaction in
145 the central metabolism by catalyzing the oxidative decarboxylation of pyruvate to
146 acetyl-CoA (Fig. 1), thereby producing one molecule CO2 and one molecule NADH
147 per mol pyruvate. Previous studies already revealed that PDHC-deficient strains
148 feature an about two-fold decreased excretion of CO2 during exponential growth (30).
- 149 It is therefore likely to assume that a reduction of the intracellular CO2/HCO3 pool
150 has an impact on metabolic flux distribution in PDHC-deficient strains. A simple
151 growth comparison of the C. glutamicum wild type and a strain lacking the ΔaceE
152 gene (thus PDHC-deficient) revealed a slightly increased lag phase of the ΔaceE
153 strain which especially became evident when cells were grown in microtiter plates,
154 while the maximal growth rate appeared to be unaffected (ΔaceE: 0.40 ± 0.01 h-1;
155 WT: 0.41 ± 0.01 h-1) (Fig. 2A). Remarkably, this lag phase was completely eliminated
- 156 by the addition of 100 mM HCO3 attributing the effect to the lowered intracellular
- 157 CO2/HCO3 pool of this mutant (Fig. 2B).
158 In C. glutamicum, the pyruvate carboxylase (PCx) encoded by pyc is the dominating
- 159 HCO3 -dependent anaplerotic enzyme required for refueling the TCA cycle via
160 oxaloacetate during growth on glycolytic carbon sources (24). In the following, we
161 deleted pyc in the ΔaceE background to examine the influence of different
- 162 CO2/HCO3 levels on bicarbonate-dependent anaplerosis. Remarkably, the additional
163 deletion of the pyc gene resulted in an elongated lag phase of about 20 hours and a
-1 164 reduced growth rate of strain ΔaceE Δpyc (µmax = 0.27 ± 0.02 h compared to 0.40 ±
165 0.01 h-1 of strain ΔaceE) during cultivation in CGXII minimal medium containing
166 glucose and acetate (Fig. 2A). An extended lag phase was also observed for the 8
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
167 Δpyc single mutant, but the lag phase was significantly less prominent compared to
168 the PDHC-deficient variant ΔaceE Δpyc (Fig. 2C). The growth defect of both strains
169 was successfully complemented by re-introducing the pyc gene on the plasmid pJC1
170 under control of its own promoters (Fig. 2 C). Remarkably, this extended lag phase
171 was only observed in the presence of glucose, but not during cultivation on minimal
172 medium containing acetate as sole carbon source (Fig. 2A).
173
174 Glucose uptake results in strongly retarded growth in strains lacking pyruvate
175 carboxylase
176 In order to verify whether glucose uptake caused the retarded growth, the ΔaceE
177 Δpyc strain was cultivated in the presence of 154 mM acetate and increasing
178 amounts of glucose (100-250 mM). In line with our hypothesis, the lag phase showed
179 a stepwise increase when higher amounts of glucose were added to the medium
180 (Fig. 3A). Consistently, ΔaceE Δpyc also showed significant elongated lag phases on
181 the PTS sugars fructose and sucrose (Fig. S1). By contrast, the glucose
182 concentration did not significantly affect the growth of the ΔaceE strain (Fig. 3A), in
183 which the major anaplerotic enzyme PCx still replenishes the oxaloacetate pool.
184 To link the observed growth phenotype to the uptake of glucose, we deleted the ptsG
185 gene, encoding the glucose specific permease of the PTS system, in the ΔaceE and
186 ΔaceE Δpyc strain background. Growth and glucose consumption rates of resulting
187 strains were analyzed during growth on glucose and acetate. Remarkably, deletion of
188 ptsG fully restored the growth of the parental strain ΔaceE Δpyc (ΔaceE Δpyc: 0.27 ±
189 0.02 h-1; ΔaceE Δpyc ΔptsG: 0.39 ± 0.02 h-1), but resulted in reduced final
190 backscatter values comparable to growth on acetate alone (Fig. 3B, Table 3). Strains
191 lacking the ptsG gene consumed only minor amounts of glucose, while strains ΔaceE
192 and ΔaceE Δpyc consumed glucose after entering the stationary phase (Fig. S2). 9
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
193 Glucose uptake rates of ΔaceE Δpyc were significantly reduced to 10.87 ± 4.95 nmol
194 min-1 g-1 compared to ΔaceE with 16.81 ± 3.01 nmol min-1 g-1. In contrast, deletion of
195 ptsG did not restore growth on the PTS substrate fructose (Fig. S3), which is mainly
196 imported by the permease encoded by ptsF (31, 32). Altogether, these findings
197 clearly link the observed growth defect of the ΔaceE Δpyc strain to the uptake of
198 glucose or alternative PTS substrates.
199
- 200 Increased HCO3 availability improved anaplerotic flux in pyc mutants
- 201 In the following, we used different strategies to alter the intracellular CO2/HCO3 pool.
- 202 This was achieved by (1) supplementing the medium with HCO3 , (2) by mutation of
203 the endogenous urease gene producing CO2 from urea in the early growth phase, (3)
204 by increasing the pH of the growth medium, and (4) by re-routing metabolic flux via
205 the pentose phosphate pathway (Fig. 3).
- 206 Similar as observed for the parental strain ΔaceE, addition of 100 mM HCO3
207 eliminated the lag phase of PDHC-deficient strains lacking the pyc gene (Fig. 2B)
208 during microtiter plate cultivation on glucose-acetate mixtures. Restoration of growth
209 was not possible in a strain lacking both anaplerotic enzymes PCx and PEPCx,
- 210 indicating that the positive effect of HCO3 on the ΔaceE Δpyc strain depends on the
211 increased flux over PEPCx (Fig. 4A, Fig. S4). In contrast to PCx-deficient strains, no
212 negative impact of glucose on growth of the strain ΔaceE Δppc was observed (Fig.
213 S5) confirming again the superior role of PCx in C. glutamicum (24, 33). Interestingly,
214 we also observed a significant impact of the culture volume on growth of ΔaceE
215 Δpyc. Since efficient liquid-air mixing will continuously remove CO2 from the culture
216 medium, the lag phase of ΔaceE Δpyc was much more prominent during microtiter
217 plate cultivation (~17 h) compared to cultures in shake flasks (~7 h lag phase in 100
218 ml cultures; ~13 h lag phage in 50 ml) (Fig. S6). 10
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
219 The urease enzyme, catalyzing the degradation of urea to ammonium and carbon
- 220 dioxide, represents a further important contributor to the intracellular CO2/HCO3 pool,
221 especially in the early exponential phase when cell densities and decarboxylation
222 reaction rates are low. In a previous study, the ureD-E188* mutation was found to
223 abolish urease activity in C. glutamicum and to enhance L-valine production of
224 PDHC-deficient strains (34). In this study, we introduced the ureD-E188* mutation
225 into the ΔaceE Δpyc background to examine the influence of an even lower
- 226 CO2/HCO3 level on anaplerosis in this particular strain background. The additional
227 inactivation of the urease enzyme resulted in a significantly elongated lag phase of
- 228 about 46 hours, which could also be complemented by addition of HCO3 (Fig. 4B)
- 229 demonstrating the importance of the CO2/HCO3 pool during the initial growth phase.
230 Furthermore, we assessed the impact of urease mutation on L-valine and L-alanine
231 production of strain ΔaceE and strain ΔaceE ureD-E188*and the complementation
- 232 via addition of HCO3 . Within 28 hours of cultivation, the ΔaceE strain produced 15
233 mM L-valine and 10 mM L-alanine as by-product, while the ΔaceE ureD-E188*
- 234 accumulated 54 mM L-valine and 38 mM L-alanine (Fig. 5). Addition of HCO3 ,
235 however, significantly reduced L-valine production by 54% (ΔaceE) and 59% (ΔaceE
236 ureD-E188*), respectively. The by-product L-alanine increased by 22% for the ΔaceE
237 strain and decreased by 51% for the ΔaceE ureD-E188* strain (Fig. 5). The example
238 of L-valine/L-alanine production illustrates the important role of the intracellular
- 239 CO2/HCO3 levels on metabolic flux in PDHC-deficient strains.
- 240 Another approach to influence the extracellular HCO3 availability was to change the
241 pH conditions (2). According to the Bjerrum plot, CO2 is predominant at acidic
2- 242 conditions, while at alkaline conditions CO3 is the dominant form. At a pH of about
- - 243 8, HCO3 is the prevalent form. As HCO3 is the substrate of the anaplerotic enzymes
244 PCx and PEPCx, the pH in the cultivation medium was adjusted to pH 8 using KOH 11
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
- 245 leading to an equilibrium shift towards higher HCO3 availability. As negative control,
246 the pH was adjusted to pH 6 using HCl. While growth of both strains was retarded at
247 pH 6, it was significantly enhanced at pH 8 (Fig. 4C). Thus, elevation of the culture
248 pH improved growth of PDHC-deficient strains likely by increasing anaplerotic flux.
249 With the reaction catalyzed by the 6-phosphogluconate dehydrogenase (6PGDH),
250 the PPP comprises another decarboxylation reaction in central carbon metabolism,
- 251 thereby contributing to the intracellular CO2/HCO3 pool. Remarkably, the lag phases
252 of ΔaceE Δpyc and ΔaceE ureD-E188* Δpyc were mostly complemented during
253 growth on acetate and gluconate – the latter entering the PPP via gluconate-6-
254 phosphate (Fig. 4D). Growth on ribose, which enters the PPP via ribulose-5-
255 phosphate and thereby bypassing the decarboxylation catalyzed by 6PGDH, showed
256 a significantly elongated lag phase in pyc deficient strains (Fig. 4D). In a further
257 attempt, metabolic/glycolytic flux was re-routed through the PPP by deleting pgi,
258 encoding the glucose-6-phosphate isomerase. However, deletion of pgi resulted in a
259 severe growth defect in the ΔaceE background. In this context the growth of ΔaceE
260 Δpyc featured only a minor, non-significant improvement (Fig. S7).
261 Taken together, our findings highlight the important impact of the intracellular
- 262 CO2/HCO3 pool on metabolic flux in central carbon metabolism, which is especially
263 evident in strain ΔaceE Δpyc lacking a central decarboxylation reaction and the key
264 carboxylase PCx in C. glutamicum.
265
266 Refueling the TCA cycle improves growth of pyc mutants
267 Glucose catabolism requires sufficient anaplerotic flux to replenish TCA cycle
268 intermediates providing precursors for various anabolic pathways. Therefore, we
269 tested whether the addition of TCA intermediates would complement the negative
270 impact of glucose on the growth of ΔaceE Δpyc. Remarkably, all tested TCA 12
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
271 intermediates (succinate, malate, citrate and the glutamate containing dipeptide Glu-
272 Ala) reduced the extended lag phase of ΔaceE Δpyc during growth on glucose-
273 acetate mixtures (Fig. 6A). The dipeptide Glu-Ala and succinate reduced the
274 elongated lag phase to a high extent by 90 % (Δt = 10 %) and 85 % (Δt = 15 %),
275 respectively, while the effects of citrate and malate were weaker (Δt = 40 % and 38
276 %, respectively) (Table 4).
277 Based on previous studies, it was known that the glyoxylate shunt might be switched
278 off during growth on glucose-acetate mixtures, due to the inhibitory effect of glucose
279 (35). Here, overexpression of the genes aceA and aceB, encoding the glyoxylate
280 shunt enzymes isocitrate lyase and malate synthase, resulted in a significantly
281 shorter lag phase of strain ΔaceE Δpyc (Fig. 6B). Overexpression was accomplished
282 using the vector pJC1 harboring an inducible Ptac promoter in front of two synthetic
283 operon variants: pJC1-aceA-linker-aceB and pJC1-aceB-linker-aceA. However, it has
284 to be noted that the overexpression of aceA and aceB led to slightly longer lag
285 phases of strain ΔaceE than compared to the empty vector control (Fig. 6B).
286 Expression from the leaky Ptac promoter yielded the best result, while induction with
287 IPTG resulted in severe growth defects (data not shown).
288 Unusual intermediate accumulation or depletion can provide valuable information
289 regarding intracellular flux imbalances causing the observed growth defects. Gas
290 chromatography-time of flight (GC-ToF) analysis was performed for analysis of the
291 metabolic states by comparing samples of ΔaceE during early exponential phase,
292 ΔaceE Δpyc at the same time point during the lag phase, as well as ΔaceE Δpyc
293 during early exponential phase (see Fig. S8 showing the sampling scheme).
- 294 Additionally, samples of both PDHC-deficient strains cultivated with 100 mM HCO3
295 were measured and WT cells during exponential phase were used as reference
296 (Table S3). No significant intermediate accumulation or occurrence of unusual 13
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
297 compounds was observed in the lag phase sample of ΔaceE Δpyc compared to all
298 other samples. Although oxaloacetate cannot be measured by this method, a
299 complete depletion of oxaloacetate-derived aspartate was detected in ΔaceE Δpyc
- 300 cells during the lag phase but not during exponential phase or in HCO3
301 complemented samples. This supports the hypothesis that oxaloacetate depletion is
302 the key reason for the elongated lag phase.
303 In order to identify oxaloacetate depletion as key growth limiting factor and to
304 decouple this effect from anaplerosis-dependent replenishment, strains Δpyc, Δpyc
305 Δppc and ΔaceE Δpyc Δppc were tested in the presence of different TCA
306 intermediates. Here, mutants lacking both anaplerotic reactions (PCx and PEPCx,
307 strains Δpyc Δppc and ΔaceE Δpyc Δppc) were not able to grow on glucose acetate
308 mixtures at all, but grew on gluconeogenetic substrates like acetate (Fig. 4A, Fig.
309 S4). Again, all TCA intermediates were able to complement the glucose sensitivity of
310 these pyc-deficient strains and the growth of Δpyc Δppc and ΔaceE Δpyc Δppc was
311 effectively restored (Table 5). Nevertheless, slight differences were observed, which
312 might also be caused by differences in the uptake of these carbon sources.
313
314 Adaptation of C. glutamicum ΔaceE Δpyc to growth on glucose and acetate
315 In a previous study, Kotte et. al reported on an elongated lag phase of E. coli as a
316 result of glucose-gluconeogenetic substrate shifts (36). This phenomenon was
317 ascribed to the formation of a small subpopulation, which is able to start growing after
318 carbon source switches (36). Usually, the growth of a small subpopulation is
319 obscured by typical bulk measurements, but can be visualized by single-cell
320 approaches, such as flow cytometry. To this end, the membranes of ΔaceE and
321 ΔaceE Δpyc cells were stained using the non-toxic, green fluorescent dye PKH67
322 prior to inoculation. By cellular growth, the amount of fluorescent dye is diluted by 14
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
323 membrane synthesis leading to a decrease of single cell PKH67 fluorescence (36).
324 Stained ΔaceE and ΔaceE Δpyc cells were cultivated in minimal medium containing
325 glucose and acetate (Fig. 7A). As non-proliferating control, ΔaceE Δpyc cells were
326 incubated in glucose as sole carbon source. Using flow cytometry, membrane
327 staining was analyzed during the cultivation at the single-cell level (Fig. 7B). While
328 the mean of fluorescence of the whole ΔaceE population decreased from 2.7 × 104
329 a.u. to 7.5 × 102 a.u., the mean of the ΔaceE Δpyc population incubated in glucose
330 medium only shifted from 3.1 × 104 a.u. to 1.3 × 103 a.u. resulting from a minor
331 dilution/bleaching of fluorescence intensity during incubation. Remarkably, both
332 strains, ΔaceE and ΔaceE Δpyc, featured a rather heterogeneous adaptation
333 behavior on glucose and acetate mixtures which was apparently time delayed in
334 strain ΔaceE Δpyc reflecting the elongated lag phase. Already in the early phase of
335 cultivation, the populations split up in two subpopulations of low and high PKH67
336 fluorescence indicating that only a few cells start to proliferate, while a fraction of the
337 population was not able to grow under same conditions (37-39).
338 To study the adaptation of ΔaceE Δpyc, comparative transcriptome analysis were
339 performed using DNA microarrays. To this end, ΔaceE and ΔaceE Δpyc cells were
340 harvested at a comparable optical density during early exponential phase (Fig. S8).
341 Both strains were cultivated in CGXII media with 154 mM acetate and 222 mM
342 glucose. Under the chosen conditions, a total of 354 genes showed a significantly
343 altered mRNA level. While 121 genes were at least 1.7-fold upregulated (p-value <
344 0.05), 233 genes were 0.7-fold downregulated (p-value < 0.05) in strain ΔaceE Δpyc
345 (Table S4). An overview of expression changes of selected genes is shown in Table
346 6. Among the various changes on the level of gene expression, we observed a 0.59-
347 fold and 0.67-fold downregulation of the respective PTS-system components ptsI and
348 ptsG in strain ΔaceE Δpyc, which is in line with the decreased glucose uptake rates 15
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
349 of this strain (Table 3). In contrast, iolT1 encoding the myo-inositol transporter, which
350 is inter alia responsible for PTS-independent glucose uptake, was upregulated (40,
351 41). This is in line with the fact that the ptsG deficient strains still consumed minor
352 amounts of glucose. Interestingly, genes involved in acetate metabolism, including
353 aceA and aceB constituting the glyoxylate shunt, displayed an increased mRNA level
354 in strain ΔaceE Δpyc. This may also represent an important adaptive mechanism, as
355 the glyoxylate shunt needs to be active in ΔaceE Δpyc to compensate for the loss of
356 TCA intermediates during co-consumption of acetate and glucose as demonstrated
357 by the effect of aceA-aceB overexpression in this study (Fig. 6B). While the mRNA
358 level of genes encoding anaplerotic enzymes did not show a significant change, the
359 glycolytic genes pgi and pyk featured an about two-fold reduced level in the ΔaceE
360 Δpyc mutant. This might as well reflect an adaptive mechanism, as a downregulation
361 of pyk would reduce flux from PEP to pyruvate and reduced expression of pgi would
362 probably lead to an increased flux through the PPP contributing an additional
363 decarboxylation step.
364
365 Less is more - Adaptive laboratory evolution of ΔaceE Δpyc revealed rapid
366 inactivation of glucose uptake
367 In order to identify key mutations abolishing the lag phase of ΔaceE Δpyc, an
368 adaptive laboratory evolution (ALE) experiment was performed. In this ALE,
369 C. glutamicum ΔaceE Δpyc was grown in CGXII minimal medium containing 154 mM
370 acetate and 222 mM glucose in overall 16 repetitive-batch cultures. Remarkably,
371 already after eight inoculations a significantly shortened lag phase was observed;
372 after 10 inoculations strain ΔaceE Δpyc featured a similar lag phase as the strain
373 ΔaceE (Fig. 8A). In contrast, serial cultivation transfers in CGXII medium containing
374 solely acetate did not lead to improved growth in glucose-acetate medium indicating 16
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
375 that glucose acted as evolutionary selection pressure (Fig. S9). Genome re-
376 sequencing of three selected glucose-acetate clones presented in Fig. 8B revealed
377 mutations in the ptsI gene, encoding the EI enzyme of the PTS system, in two
378 independently evolved clones (Table 7). This is in agreement with the reduced
379 biomass formation of both clones, which apparently abolished glucose uptake to
380 optimize growth on acetate. The third clone showed only a slightly accelerated
381 growth, but was not impaired in biomass formation. Here, sequencing revealed a
382 mutation in the rpsC gene, encoding the ribosomal S3 protein.
383 In summary, it was indeed possible to eliminate the observed lag phase of strain
384 ΔaceE Δpyc ascribed to glucose sensitivity by an adaptive laboratory evolution
385 approach. Especially, the mutations identified in the ptsI gene are in agreement with
386 other experiments emphasizing the high selective pressure exerted by glucose on
- 387 PCx-deficient strains when the intracellular HCO3 levels are limiting.
388
389
17
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
390 Discussion
391 During growth on glucose, the pyruvate dehydrogenase complex catalyzes a central
- 392 metabolic reaction contributing to the intracellular CO2/HCO3 pool. In this study, we
- 393 systematically perturbed the intracellular CO2/HCO3 pool of PDHC-deficient
394 C. glutamicum strains to monitor the impact on growth and anaplerotic flux. Even the
395 PDHC-deficient strain C. glutamicum ΔaceE features a slightly elongated lag phase
396 on glucose-acetate mixtures compared to the wild type, while no difference in growth
397 is observed when strains are growing on gluconeogenetic substrates, like acetate.
398 This is remarkable since this strain catabolizes only a minor fraction of the provided
399 glucose in the exponential growth phase, but rather turns to the production of L-
400 valine and L-alanine from glucose in the stationary phase (11, 12). The growth defect
401 of C. glutamicum ΔaceE was, however, successfully complemented by the addition of
- 402 HCO3 which had no significant effect on growth of wild type cells, therefore, hinting
403 towards problems caused by an impaired anaplerotic flux in this strain background.
404 By comparing the CO2 production rates of the C. glutamicum ΔaceE strain and the
405 wild type, Bartek et al. could show in a previous study, that strain ΔaceE excretes
-1 -1 -1 406 only around 0.83 mmol g h of CO2, while the wild type excretes ca. 1.65 mmol g
-1 407 h CO2 (30). The lower CO2 production rates of the PDHC-deficient strain already
- 408 indicate a significant impact on the intracellular CO2/HCO3 pool affecting metabolic
- 409 flux distribution especially at low cell densities when CO2/HCO3 is limiting.
410 The anaplerotic node comprises the essential link between
411 glycolysis/gluconeogenesis and the TCA cycle (21). In contrast to most other
412 organisms, C. glutamicum possesses both anaplerotic carboxylases, PEPCx
413 (encoded by ppc) and PCx (encoded by pyc) catalyzing the anaplerotic bicarbonate-
414 dependent reactions to yield oxaloacetate from PEP or pyruvate, respectively. In C.
415 glutamicum, it was shown that these two enzymes can replace each other to a 18
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
416 certain extent depending on the intracellular concentrations of the respective
417 effectors for each enzyme. However, during growth on glucose, the biotin-containing
418 PCx contributes to the main anaplerotic activity of 90% compared to PEPCx (24, 33).
419 In our study, deletion of the pyc gene resulted in a severely elongated lag phase (>15
420 hours) of ΔaceE Δpyc strains when glucose was present in the medium. Retarded
421 growth of Δpyc strains have been reported previously and, in line with our data,
422 Blombach and co-workers also reported a severe growth defect of this strain under
- 423 low CO2/HCO3 levels (24, 29). This effect appeared to be even more severe when
424 the cells were grown in microtiter plates under CO2-stripping conditions (low culture
425 volume, high air mixing) compared to growth in higher culture volumes (Fig. S6). The
426 PDHC-deficient mutant strain C. glutamicum ΔaceE Δpyc was found to react very
427 sensitively towards small changes in the intracellular bicarbonate availability and was
428 therefore used as a test platform to systematically assess the impact of perturbations
- - 429 affecting the intracellular CO2/HCO3 pool. While the addition of HCO3 , the increase
430 of the pH as well as higher culture volumes rescued the strain, mutation of the urease
- 431 accessory protein (ureD-E188*), which lowers the intracellular CO2/HCO3 pool,
432 resulted in an even more severe phenotype. It was not possible to restore growth if
433 the second anaplerotic gene ppc was deleted as well, confirming that the
434 abovementioned measures fostered the flux over the remaining anaplerotic reaction
435 catalyzed by PEPCx.
436 The mutation in ureD was revealed by a previous biosensor-driven evolution
437 approach selecting for mutations enhancing L-valine production in C. glutamicum
438 (34). Here, inactivation of the urease enzyme by the mutation ureD-E188* reduced
439 the anaplerotic flux via PCx resulting in an increased precursor availability of
440 pyruvate-derived products like L-valine and L-alanine (34). In the present study, we
- 441 further confirmed this finding by complementation with HCO3 counteracting the effect 19
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
- 442 of the ureD-E188* mutation. In another study by Blombach et al., lower CO2/HCO3
443 levels also triggered enhanced production of L-valine and L-alanine of C. glutamicum
444 during shake flask cultivation (29). However, a positive impact of anaplerotic
445 reactions has been confirmed for other, for example, oxaloacetate-derived products
446 like L-lysine. Here, attempts to overexpress pyc or the introduction of deregulated
447 variants significantly improved L-lysine production (42, 43). However, in spite of great
448 efforts and successes in the development of L-lysine strains, the impact of altered
- 449 CO2/HCO3 levels has not been systematically assessed so far.
450 Residual glucose consumption of PDHC-deficient C. glutamicum strains has already
451 been reported in previous studies (11, 12, 30). The results of this work, however,
452 emphasize that mutants lacking the major anaplerotic route via PCx are under strong
453 evolutionary pressure in the presence of glucose. Although strains would be capable
454 to grow on the acetate supplied in the growth medium, increasing levels of glucose
455 resulted in a severely elongated lag phase (up to >40 hours with 250 mM) of C.
456 glutamicum ΔaceE Δpyc. Besides the abovementioned efforts to increase the
- 457 CO2/HCO3 pool, also the deletion of the ptsG gene itself effectively restored growth
458 on glucose, but not on other PTS substrates. This indicates that this strain is
459 sensitive to glucose, but does not have issues with regulatory functions of PtsG
460 known from e.g. E. coli (44, 45).
461 Furthermore, among the three key mutations identified in the ALE experiment, two
462 SNPs obtained from independent cell lines were located within the ptsI gene,
463 encoding the EI component of the PTS system. These findings clearly highlight the
464 problems of residual glucose consumption in strains featuring an impaired anaplerotic
465 flux. One reason for the observed growth phenotype might be stress resulting from
466 the intracellular accumulation of sugar phosphates. In Escherichia coli, sugar
467 phosphate stress can be caused by accumulation of any sugar phosphates due to a 20
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
468 block in glycolysis (46), e.g. by mutations in pgi or pfkA (47), or due to excessive
469 glucose uptake caused e.g. by overexpression of uhpT encoding a sugar phosphate
470 permease (48). The resulting metabolic imbalance causes growth inhibition. For
471 example, in C. glutamicum, accumulation of several sugar phosphates, like fructose-
472 1,6-bisphosphate or PEP, inhibits the isocitrate lyase (35). To counteract this stress,
473 E. coli triggers SgrR which consequently activates SgrS by an unknown signal. The
474 transcription factor SgrS then prevents further uptake of glucose by downregulation
475 of ptsG (46, 49, 50). This is in line with the finding of our study, showing a slight
476 downregulation of ptsG upon resumed growth of strain ΔaceE Δpyc (Table 3). This
477 reduction of glucose uptake via the PTS system which converts PEP to pyruvate
478 during glucose transport also contributes to an increase in PEP availability for the
479 remaining anaplerotic reaction catalyzed by PEPCx. Further, this is in line with an
480 upregulation of iolT1, which ensures minor usage of glucose without PEP depletion
481 (40, 41). In this context, also the downregulation of the pyk gene in strain ΔaceE
482 Δpyc can be interpreted as a potential adaptation to increase the PEP pool fostering
483 anaplerotic flux.
484 During growth on gluconeogenic carbon sources - like acetate - the transcriptional
485 repressor SugR represses the expression of the PTS genes in C. glutamicum,
486 including the glucose-specific ptsG gene, but also ptsF, ptsS and general
487 components as ptsI and ptsH (51-53). Derepression appears to be triggered by the
488 accumulation of fructose-6-phosphate (F6P), which is generated from glucose 6-
489 phosphate entering glycolysis. Consequently, F6P accumulation leads to an increase
490 of glucose consumption rates resulting in parallel catabolization of glucose and
491 acetate (15). The fact that the PDHC-deficient strains are also subject to regulation of
492 SugR was demonstrated by studies showing that ΔaceE ΔsugR strain leads to up to
493 4-fold higher glucose consumption rates (54). 21
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
494 In aerobic glucose-based bioprocesses the endogenous production of CO2 is
495 typically sufficiently high to promote microbial growth even at low cell densities.
496 However, it is a well-known fact that anaerobic growth of some bacteria, like E. coli,
- 497 requires the external supply of CO2/HCO3 to avoid long lag phases (26, 55, 56). The
- 498 critical impact of CO2/HCO3 levels may become especially evident under conditions
499 when the endogenous CO2 production rate becomes limiting. This was nicely
500 demonstrated by a recent study of Bracher et al. who could show that long lag
501 phases of engineered, but non-evolved Saccaromyces cerevisiae strains during
502 xylose fermentation could be avoided by sparging the bioreactor cultures with CO2/N2
503 mixtures (57). Alternatively, the addition of L-aspartate, whose transamination
504 provides oxaloacetate refueling the TCA, completely abolished the long adaptation
505 phase of the respective yeast strains. In line with these findings, a recent 13C-flux
506 analysis with E. coli revealed that a considerable turnover of lipids via β-oxidation
- 507 appears to be required for growth on xylose to enhance the intracellular CO2/HCO3
508 pool to a growth-promoting level enabling anaplerotic flux (58). These findings are in
509 nice agreement with our study showing that the elongated lag phase of strain C.
510 glutamicum ΔaceE Δpyc can be eliminated by different measures enhancing either
- 511 the intracellular CO2/HCO3 level or by refueling of the TCA cycle with different
512 intermediates, like citrate, malate or succinate. Overall, oxaloacetate depletion
513 appeared to represent a key issue causing the delayed growth of the ΔaceE Δpyc
514 strain. Our data revealed that this is caused by an impaired anaplerotic flux on
- 515 glucose due to low intracellular CO2/HCO3 levels as well as reduced activities of
516 glyoxylate enzymes.
517 Altogether, these results emphasize the important impact of the intracellular
- 518 CO2/HCO3 pool for metabolic flux distribution, which is especially relevant in
519 engineered strains suffering from lower endogenous CO2 production rates as 22
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
520 exemplified by PDHC-deficient strains in this study, but also by the performance of
521 pentose-fermenting yeast and E. coli strains (57, 58). Consequently, the lack of an
522 important by-product, like CO2 released by the PDHC complex, may have a
523 significant impact on cellular metabolism and growth, especially on glycolytic
524 substrates demanding a high flux via the anaplerotic reactions.
525
23
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
526 Material and methods
527 Bacterial strains and growth conditions
528 All bacterial strains and plasmids used in this study are listed in Table 1. Mutant
529 strains are based on the Corynebacterium glutamicum ATCC 13032 wild type strain
530 (59).
531 Standard cultivation of C. glutamicum ΔaceE cells and derivatives were performed on
532 BHI (brain heart infusion, Difco BHI, BD, Heidelberg, Germany) agar plates
533 containing 51 mM potassium acetate (hereinafter named acetate) (ChemSolute, Th.
534 Geyer, Stuttgart, Germany) at 30°C for two days. One single colony was picked and
535 incubated for 8 to 10 h at 30°C in either 4.5 ml or 1 ml BHI containing 154mM acetate
536 in reaction tubes or Deepwell-plates (VWR International, Pennsylvania, USA),
537 respectively. First pre-cultures were used to inoculate second pre-cultures in CGXII
538 minimal medium (60) supplemented with 154 mM acetate either as 10 ml culture in
539 shake flasks or in 1 ml volume in Deepwell-plates. After overnight growth, a main
540 culture was inoculated at an OD600 of 1 in CGXII media containing 154 mM acetate
541 and either 222 mM D(+)-glucose monohydrate (Riedel-de Haën, Seelze, Germany)
542 (further referred to as glucose), or any other carbon source as stated, e.g. D(-)-
543 fructose (Sigma-Aldrich, Taufkirchen, Germany) (referred to as fructose), D(+)-
544 sucrose (Roth, Karlsruhe, Germany) (referred to as sucrose), D-gluconic acid sodium
545 salt (Sigma-Aldrich, Taufkirchen, Germany) (referred to as gluconate), D(-)ribose
546 (Sigma-Aldrich, Taufkirchen, Germany) (referred to as ribose), citric acid
547 monohydrate (Merck Millipore, Darmstadt, Germany) (referred to as citrate), H-Glu-
548 Ala-OH (Bachem AG, Bubendorf, Germany) (referred to as Glu-Ala), succinic acid
549 (Sigma-Aldrich, Taufkirchen, Germany) (referred to as succinate) or L-malic acid
550 (Merck Millipore, Darmstadt, Germany) (referred to as malate). For elevating the
- 551 extracellular availability of bicarbonate, 100 mM potassium HCO3 (Merck Millipore, 24
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
552 Darmstadt, Germany) was added to the CGXII-basis solution, which was sterile-
553 filtered before further use. In order to analyze the effect of different pH levels, either
554 HCl for lowering or KOH for increasing the pH was added to the CGXII-basis, which
555 was sterile-filtered afterwards. For cultivations in the presence of TCA cycle-
556 filling/refueling substrates, 50 mM citric acid monohydrate (citrate), 3 mM H-Glu-Ala-
557 OH dipeptide (Glu-Ala), 100 mM succinic acid (succinate) or 100 mM L-malic acid
558 (malate) was used. In experiments where gluconate or ribose were used as carbon
559 source, 100 mM D-gluconic acid sodium salt or 100 mM D-ribose were added,
560 respectively. Biomass formation was monitored during cultivation in shake flasks by
561 measuring OD600 or by measuring backscatter values during microtiter plate
562 cultivation (section 2.2). Where necessary, 25 µg/ml kanamycin was also
563 supplemented.
564 For the purpose of cloning and plasmid isolations, Escherichia coli DH5α strains were
565 grown in shake flasks in Lysogeny broth (LB) medium at 37°C directly inoculated
566 from a glycerol stock or from an LB-agar plate in shake flasks at 37°C. If necessary
567 for selection, 50 µg/ml kanamycin was also supplemented.
568 Microtiter plate cultivation
569 Online monitoring of growth and/or pH was performed in 48-well microtiter Flower
570 Plates® (m2p-labs GmbH, Baesweiler, Germany) sealed with sterile breathable
571 rayon films (VWR International, Pennsylvania, US) in a BioLector® microtiter
572 cultivation system (m2p-labs GmbH, Baesweiler, Germany) (61). The cultivation
573 conditions were adjusted as described earlier (62) and biomass formation was
574 recorded each 15 min as the backscattered light intensity (light wavelength 620 nm;
575 signal gain factor of 20) for 24 to 72 h at 30°C and 1,200 rpm. pH measurements
576 were performed with 48-well microtiter Flower Plates® being equipped with pH
577 optodes. Obtained data was evaluated using the software BioLection (m2p-labs, 25
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
578 Baesweiler, Germany) and Graph Pad Prism (Graphpad Prism 7 Software, Inc.,
579 California, USA).
580 Adaptive laboratory evolution
581 The adaptive laboratory evolution (ALE) of C. glutamicum ΔaceE and ΔaceE Δpyc
582 was performed in Deepwell-plates (VWR International, Pennsylvania, US) in a main
583 culture of 1 ml CGXII media containing 154 mM acetate and 222 mM glucose or
584 solely 154 mM acetate (as control without selection pressure) adjusted to an OD600 of
585 1. Cells were cultivated for 2-3 days before the next generation was inoculated
586 starting at an OD600 of 1 and cultivated again for 2-3 days. After each generation
587 step, glycerol stocks of cultures were prepared (20% glycerol) and stored at -80°C
588 allowing growth analysis and DNA sequencing of individual inoculations. In total, 16
589 serial transfers were analyzed.
590 Cloning techniques and recombinant DNA work
591 Standard molecular biology methods were performed according to J. Sambrook and
592 D. W. Russell (63). C. glutamicum ATCC 13032 chromosomal DNA was used as
593 template for PCR amplification of DNA fragments and prepared as described earlier
594 (64). DNA fragment and plasmid sequencing as well as synthesis of oligonucleotides
595 was performed by Eurofins Genomics (Ebersberg, Germany).
596 For the construction of plasmids (Table S2), DNA fragments were amplified using the
597 respective oligonucleotides (Table S1) and enzymatically assembled into a vector
598 backbone according to (65).
599 To achieve genomic deletion of pyc, ptsG, ppc and pgi, two-step homologous
600 recombination using the pK19mobsacB-system (66) was implemented. The suicide
601 plasmids (compare Table 2 and Table S2) were isolated from E. coli cells using the
602 QIAprep spin miniprep kit (Qiagen, Hilden, Germany). Electrocompetent C.
603 glutamicum ΔaceE and ΔaceE Δpyc cells were transformed with these plasmids by 26
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
604 electroporation (67). The first and second recombination events were performed and
605 verified as described in previous studies (68). The deletion of pyc, ptsG, ppc and pgi
606 was reviewed by amplification and sequencing using primers shown in Table S1.
607 Measurement of glucose concentrations
608 To measure the glucose concentration of the culture media at different time points,
609 cultivation was performed in 50 ml in shaking flasks. During cultivation, 0.5 ml
610 samples were taken every 3 h and centrifuged (16,000 x g). Supernatant was
611 collected and stored at -20°C until use.
612 The measurement of the actual glucose concentration was performed using D-
613 Glucose UV-Test Kit (r-biopharm, Darmstadt, Germany). Performance and
614 calculations according to manufacturer’s instructions were done, while absorption
615 was measured at 340 nm.
616 Further calculations concerning the glucose uptake rates could be done based on the
617 obtained glucose concentrations and OD600 values. According to (18), Formula 1 was
618 used:
619
620 Formula 1: Determination of glucose concentration in nmol gDW-1 h-1
푆 푛푚표푙 ∗ 푙−1 ∗ 푂퐷 −1 푛푚표푙 ( ) ∗ µ [( 600 ) ∗ ℎ−1] = [ ] −1 −1 푀 𝑔퐷푊 ∗ 푙 ∗ 푂퐷600 𝑔퐷푊 ∗ ℎ
621
-1 622 While S represents the slope of the glucose concentration versus OD600 [nmol * l *
-1 -1 - 623 OD600 ], M is the correlation between dry weight and OD600 [g dry weight * l * OD600
1 -1 624 ], and µ is the growth rate [h ]. According to A. Kabus et al. (69), an OD600 of 1
625 corresponds to 0.25 g dry weight l-1, so that this value was used as M throughout this
626 calculations.
27
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
627 Quantification of amino acid production
628 Using ultra-high performance liquid chromatography (uHPLC), amino acids were
629 quantified as ortho-phthaldialdehyde derivatives by automatic pre-column
630 derivatization. Separation of derivatives by reverse-phase chromatography was
631 performed on an Agilent 1290 Infinity LC ChemStation (Agilent, Santa Clara, USA)
632 equipped with a fluorescence detector. As eluent for the Zorbax Eclipse AAA 3.5
633 micron 4.6 x 7.5 mm column (Agilent, Santa Clara, USA), a gradient of Na-borate
634 buffer (10 mM Na2HPO4; 10 mM Na2B4O7, pH 8.2; adapted to operater’s guide) and
635 methanol was applied. Prior to analysis, samples were centrifuged for 10 min at
636 16,000 x g and 4°C and diluted 1:100.
637 Monitoring of cellular proliferation by cell staining
638 For staining of proliferating cells, the PKH67 green fluorescent cell linker kit for
639 general cell membrane labeling (Sigma Aldrich, Munich, Germany) was used and the
640 protocol was adapted to O. Kotte et al. (36). From an exponentially growing pre-
641 culture in CGXII minimal medium containing 222 mM glucose and 154 mM acetate,
642 1.5 x 109 cells were harvested by centrifugation for four minutes at 4000 g and 4°C.
643 Then, the cells were washed again in 5 ml ice-cold CGXII basic solution, without
644 MgSO4, CaCl2, biotin, trace elements and protocatechuic acid. For staining, the cell
645 pellet was resuspended in 500 µl dilution buffer C (Sigma Aldrich, Munich, Germany)
646 at room temperature and a freshly prepared mixture of 10 µl PKH67 dye (Sigma
647 Aldrich, Munich, Germany) and 500 µl dilution buffer C was added. Subsequently,
648 cells were incubated for three minutes at room temperature and afterwards, 4 ml ice-
649 cold filtered CGXII basic solution containing 1% (w/v) bovine serum albumin, 1 mM
650 MgSO4 and 0.1 mM CaCl2 were added. Then, the cells were centrifuged for four
651 minutes at 4000 g and 4°C and the cell pellet was washed twice. Finally, the cells
28
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
652 were resuspended in CGXII minimal medium containing 222 mM glucose and 154
653 mM acetate and cultivated in a microtiter cultivation system.
654 Flow cytometry
655 Flow cytometric (FC) analyzes were conducted on a FACSAria II flow cytometer
656 (Becton Dickinson, San Jose, USA) equipped with a blue solid state laser (488 nm
657 excitation). Forward-scatter characteristics (FSC) and side-scatter characteristics
658 (SSC) were recorded as small-angle and orthogonal scatters of the 488-nm laser,
659 respectively. PKH67 fluorescence was detected using a 502-nm long-pass and a
660 530/30-nm band-pass filter set. FACS-Diva software 6.0 was used to monitor the
661 measurements. During analyzes, thresholding on FSC was applied to remove
662 background noise. For FC analyzes, PKH67-stained culture samples were diluted to
663 an OD600 of 0.05 in FACSFlow™ sheath fluid buffer (BD, Heidelberg, Germany). The
664 analysis software FlowJo V.10.0.8 was used to visualize and evaluate the data (Tree
665 Star, Ashland, USA).
666 DNA microarrays
667 For analysis of the transcriptome, C. glutamicum ΔaceE and ΔaceE Δpyc were
668 cultivated in triplicates as described in 2.1 in 50 ml CGXII containing 154 mM acetate
669 and 222 mM glucose in shake flasks. After reaching exponential phase at an OD600
670 of around 12-15, the cell suspension was harvested by centrifugation (4256 g, 10
671 min, 4°C). The resulting pellet was directly frozen in liquid nitrogen and stored at -
672 80°C. RNA preparation, cDNA synthesis as well as microchip hybridization, scanning
673 and evaluation were performed as described by previous studies (70).
674 GC-ToF-MS analysis
675 For analysis of the metabolome, samples of C. glutamicum ΔaceE in exponential
676 phase as well as ΔaceE Δpyc in stationary and exponential phase were taken and
677 cell disruption was accomplished using hot methanol. Further sample preparation, 29
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
678 derivatization, MS operation and peak identification was accomplished according to
679 N. Paczia et al. (71) in an Agilent 6890N gas chromatograph (Agilent, Santa Clara,
680 USA) coupled to a Waters Micromass GCT Premier high resolution time-of-flight
681 mass spectrometer (Waters, Milford, USA). Identification of known metabolites was
682 accomplished using in-house database JuPoD, the commercial database NIST17
683 (National Institute of Standards and Technology, US) and the database GMD (MPI of
684 Molecular Plant Physiology, Golm) (72).
685 Whole-genome sequencing
686 In order to sequence the whole genome of C. glutamicum ΔaceE Δpyc mutants from
687 the ALE experiment using next-generation sequencing (NGS), genomic DNA was
688 prepared using the NucleoSpin® Microbial DNA Kit (Macherey-Nagel, Düren,
689 Germany) according to manufacturer’s instructions. Consequently, the concentrations
690 of the purified genomic DNA were measured using Qubit Fluorometer 2.0 (Invitrogen,
691 California, USA) according to manufacturer’s instructions. Overall, 4 µg purified
692 genomic DNA was employed for the preparation of genome sequencing using
693 TruSeq® DNA Library Prep Kit and MiSeq Reagent Kit v1 (Illumia, San Diego, USA)
694 according to manufacturer’s instructions. The sequencing run was performed in a
695 MiSeq system (Illumina, San Diego, USA). Data analysis and base calling was
696 performed with the Illumina instrument software. Consequent fastq output files were
697 examined using CLC Genomics Workbench 9 (Qiagen Aarhus, Denmark).
698
699 Acknowledgements
700 We thank Jochem Gätgens for performing GC-ToF analyses, as well as Bastian
701 Blombach and Stephan Noack for fruitful discussions. The authors acknowledge
702 financial support by the Helmholtz Association (grant W2/W3-096).
703 30
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
704 References
705 1. Bailey JE, Ollis DF. 1986. Biochemical Engineering Fundamentals. NY:
706 McGraw-Hill, New York.
707 2. Blombach B, Takors R. 2015. CO2 – Intrinsic Product, Essential Substrate,
708 and Regulatory Trigger of Microbial and Mammalian Production Processes.
709 Front Bioeng Biotechnol 3.
710 3. de Kok A, Hengeveld AF, Martin A, Westphal AH. 1998. The pyruvate
711 dehydrogenase multi-enzyme complex from Gram-negative bacteria. Biochim
712 Biophys 1385:353-366.
713 4. Eikmanns BJ, Blombach B. 2014. The pyruvate dehydrogenase complex of
714 Corynebacterium glutamicum: an attractive target for metabolic engineering. J
715 Biotechnol 192 Pt B:339-45.
716 5. Schreiner ME, Fiur D, Holátko J, Pátek M, Eikmanns BJ. 2005. E1 Enzyme of
717 the Pyruvate Dehydrogenase Complex in Corynebacterium glutamicum :
718 Molecular Analysis of the Gene and Phylogenetic Aspects. J Bacteriol
719 187:6005-6018.
720 6. Becker J, Rohles CM, Wittmann C. 2018. Metabolically engineered
721 Corynebacterium glutamicum for bio-based production of chemicals, fuels,
722 materials, and healthcare products. Metab Eng 50:122-141.
723 7. Kogure T, Inui M. 2018. Recent advances in metabolic engineering of
724 Corynebacterium glutamicum for bioproduction of value-added aromatic
725 chemicals and natural products. Appl Microbiol Biotechnol 102:8685-8705.
726 8. Wang X, Zhang H, Quinn PJ. 2018. Production of L-valine from metabolically
727 engineered Corynebacterium glutamicum. Appl Microbiol Biotechnol
728 102:4319-4330.
31
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
729 9. Wendisch VF, Mindt M, Perez-Garcia F. 2018. Biotechnological production of
730 mono- and diamines using bacteria: recent progress, applications, and
731 perspectives. Appl Microbiol Biotechnol 102:3583-3594.
732 10. Eggeling L, Bott M. 2005. Handbook of Corynebacterium glutamicum.
733 Academic Press, Inc., Boca Raton, FL.
734 11. Blombach B, Schreiner ME, Bartek T, Oldiges M, Eikmanns BJ. 2008.
735 Corynebacterium glutamicum tailored for high-yield L-valine production. Appl
736 Microbiol Biotechnol 79:471-9.
737 12. Blombach B, Schreiner ME, Holatko J, Bartek T, Oldiges M, Eikmanns BJ.
738 2007. L-valine production with pyruvate dehydrogenase complex-deficient
739 Corynebacterium glutamicum. Appl Environ Microbiol 73:2079-84.
740 13. Buchholz J, Schwentner A, Brunnenkan B, Gabris C, Grimm S, Gerstmeir R,
741 Takors R, Eikmanns BJ, Blombach B. 2013. Platform engineering of
742 Corynebacterium glutamicum with reduced pyruvate dehydrogenase complex
743 activity for improved production of L-lysine, L-valine, and 2-ketoisovalerate.
744 Appl Environ Microbiol 79:5566-75.
745 14. Cocaign M, Monnet C, Lindley ND. 1993. Batch kinetics of Corynebacterium
746 glutamicum during growth on various carbon substrates: use of substrate
747 mixtures to localise metabolic bottlenecks. Appl Microbiol Biotechnol 40:526-
748 530.
749 15. Wendisch VF, de Graaf AA, Sahm H, Eikmanns BJ. 2000. Quantitative
750 Determination of Metabolic Fluxes during Coutilization of Two Carbon
751 Sources: Comparative Analyses with Corynebacterium glutamicum during
752 Growth on Acetate and/or Glucose. J Bacteriol 182:3088-3096.
32
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
753 16. Dominguez H, Cocaign-Bousquet M, Lindley ND. 1997. Simultaneous
754 consumption of glucose and fructose from sugar mixtures during batch growth
755 of Corynebacterium glutamicum. Appl Microbiol Biotechnol 47:600-603.
756 17. Stansen C, Uy D, Delaunay S, Eggeling L, Goergen J-L, Wendisch VF. 2005.
757 Characterization of a Corynebacterium glutamicum lactate utilization operon
758 induced during temperature-triggered glutamate production. Appl Environ
759 Microbiol 71:5920-5928.
760 18. Frunzke J, Engels V, Hasenbein S, Gätgens C, Bott M. 2008. Co-ordinated
761 regulation of gluconate catabolism and glucose uptake in Corynebacterium
762 glutamicum by two functionally equivalent transcriptional regulators, GntR1
763 and GntR2. Mol Microbiol 67:305-322.
764 19. Cozzone AJ. 1998. Regulation of Acetate Metabolism by Protein
765 Phosphorylation in Enteric Bacteria. Annu Rev Microbiol 52:127-164.
766 20. Cronan J, P. Clark D. 2005. Two-Carbon Compounds and Fatty Acids as
767 Carbon Sources. EcoSal Plus 1.
768 21. Sauer U, Eikmanns BJ. 2005. The PEP-pyruvate-oxaloacetate node as the
769 switch point for carbon flux distribution in bacteria. FEMS Microbiol Rev
770 29:765-94.
771 22. Eikmanns BJ, Follettie MT, Griot MU, Sinskey AJ. 1989. The
772 phosphoenolpyruvate carboxylase gene of Corynebacterium glutamicum:
773 Molecular cloning, nucleotide sequence, and expression. Mol Gen Genet
774 218:330-339.
775 23. O'Regan M, Thierbach G, Bachmann B, Villeval D, Lepage P, Viret JF,
776 Lemoine Y. 1989. Cloning and nucleotide sequence of the
777 phosphoenolpyruvate carboxylase-coding gene of Corynebacterium
778 glutamicum ATCC13032. Gene 77:237-251. 33
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
779 24. Peters-Wendisch PG, Kreutzer C, Kalinowski J, Patek M, Sahm H, Eikmanns
780 BJ. 1998. Pyruvate carboxylase from Corynebacterium glutamicum:
781 characterization, expression and inactivation of the pyc gene. Microbiology
782 144 (Pt 4):915-27.
783 25. Kai Y, Matsumura H, Inoue T, Terada K, Nagara Y, Yoshinaga T, Kihara A,
784 Tsumura K, Izui K. 1999. Three-dimensional structure of phosphoenolpyruvate
785 carboxylase: a proposed mechanism for allosteric inhibition. Proceedings of
786 the National Academy of Sciences of the United States of America 96:823-
787 828.
788 26. Repaske R, Repaske AC, Mayer RD. 1974. Carbon dioxide control of lag
789 period and growth of Streptococcus sanguis. J Bacteriol 117:652-659.
790 27. Talley RS, Baugh CL. 1975. Effects of bicarbonate on growth of Neisseria
791 gonorrhoeae: replacement of gaseous CO2 atmosphere. Appl Microbiol
792 29:469-471.
793 28. Bäumchen C, Knoll A, Husemann B, Seletzky J, Maier B, Dietrich C,
794 Amoabediny G, Buchs J. 2007. Effect of elevated dissolved carbon dioxide
795 concentrations on growth of Corynebacterium glutamicum on D-glucose and
796 L-lactate. J Biotechnol 128:868-74.
797 29. Blombach B, Buchholz J, Busche T, Kalinowski J, Takors R. 2013. Impact of
- 798 different CO2/HCO3 levels on metabolism and regulation in Corynebacterium
799 glutamicum. J Biotechnol 168:331-40.
800 30. Bartek T, Blombach B, Lang S, Eikmanns BJ, Wiechert W, Oldiges M, Noh K,
801 Noack S. 2011. Comparative 13C metabolic flux analysis of pyruvate
802 dehydrogenase complex-deficient, L-valine-producing Corynebacterium
803 glutamicum. Appl Environ Microbiol 77:6644-52.
34
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
804 31. Dominguez H, Lindley ND. 1996. Complete Sucrose Metabolism Requires
805 Fructose Phosphotransferase Activity in Corynebacterium glutamicum To
806 Ensure Phosphorylation of Liberated Fructose. Appl Environ Microbiol
807 62:3878-80.
808 32. Moon MW, Kim HJ, Oh TK, Shin CS, Lee JS, Kim SJ, Lee JK. 2005. Analyses
809 of enzyme II gene mutants for sugar transport and heterologous expression of
810 fructokinase gene in Corynebacterium glutamicum ATCC 13032. FEMS
811 Microbiol Lett 244:259-66.
812 33. Peters-Wendisch PG, Eikmanns BJ, Thierbach G, Bachmann B, Sahm H.
813 1993. Phosphoenolpyruvate carboxylase in Corynebacterium glutamicum is
814 dispensable for growth and lysine production. FEMS Microbiol Lett 112:269-
815 274.
816 34. Mahr R, Gätgens C, Gätgens J, Polen T, Kalinowski J, Frunzke J. 2015.
817 Biosensor-driven adaptive laboratory evolution of l-valine production in
818 Corynebacterium glutamicum. Metab Eng 32:184-94.
819 35. Reinscheid DJ, Eikmanns BJ, Sahm H. 1994. Characterization of the isocitrate
820 lyase gene from Corynebacterium glutamicum and biochemical analysis of the
821 enzyme. J Bacteriol 176:3474-3483.
822 36. Kotte O, Volkmer B, Radzikowski JL, Heinemann M. 2014. Phenotypic
823 bistability in Escherichia coli's central carbon metabolism. Mol Syst Biol
824 10:736.
825 37. Takhaveev V, Heinemann M. 2018. Metabolic heterogeneity in clonal microbial
826 populations. Curr Opin Microbiol 45:30-38.
827 38. Heinemann M, Zenobi R. 2011. Single cell metabolomics. Curr Opin
828 Biotechnol 22:26-31.
35
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
829 39. Davis KM, Isberg RR. 2016. Defining heterogeneity within bacterial
830 populations via single cell approaches. Bioessays 38:782-90.
831 40. Ikeda M, Mizuno Y, Awane SI, Hayashi M, Mitsuhashi S, Takeno S. 2011.
832 Identification and application of a different glucose uptake system that
833 functions as an alternative to the phosphotransferase system in
834 Corynebacterium glutamicum. Appl Microbiol Biotechnol 90:1443-1451.
835 41. Lindner SN, Seibold GM, Henrich A, Krämer R, Wendisch VF. 2011.
836 Phosphotransferase system-independent glucose utilization in
837 Corynebacterium glutamicum by inositol permeases and glucokinases. Appl
838 Environ Microbiol 77:3571-3581.
839 42. Peters-Wendisch PG, Schiel B, Wendisch VF, Katsoulidis E, Mockel B, Sahm
840 H, Eikmanns BJ. 2001. Pyruvate carboxylase is a major bottleneck for
841 glutamate and lysine production by Corynebacterium glutamicum. J Mol
842 Microbiol Biotechnol 3:295-300.
843 43. Kortmann M, Mack C, Baumgart M, Bott M. 2019. Pyruvate Carboxylase
844 Variants Enabling Improved Lysine Production from Glucose Identified by
845 Biosensor-Based High-Throughput Fluorescence-Activated Cell Sorting
846 Screening. ACS Synth Biol 8:274-281.
847 44. Chatterjee R, Millard CS, Champion K, Clark DP, Donnelly MI. 2001. Mutation
848 of the ptsG gene results in increased production of succinate in fermentation
849 of glucose by Escherichia coli. Appl Environ Microbiol 67:148-154.
850 45. Wang Q, Wu C, Chen T, Chen X, Zhao X. 2006. Expression of galactose
851 permease and pyruvate carboxylase in Escherichia coli ptsG mutant increases
852 the growth rate and succinate yield under anaerobic conditions. Biotechnol
853 Lett 28:89-93.
36
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
854 46. Vanderpool CK, Gottesman S. 2007. The novel transcription factor SgrR
855 coordinates the response to glucose-phosphate stress. J Bacteriol 189:2238-
856 48.
857 47. Morita T, El-Kazzaz W, Tanaka Y, Inada T, Aiba H. 2003. Accumulation of
858 glucose 6-phosphate or fructose 6-phosphate is responsible for destabilization
859 of glucose transporter mRNA in Escherichia coli. J Biol Chem 278:15608-14.
860 48. Kadner RJ, Murphy GP, Stephens CM. 1992. Two mechanisms for growth
861 inhibition by elevated transport of sugar phosphates in Escherichia coli. J Gen
862 Microbiol 138:2007-14.
863 49. Kessler JR, Cobe BL, Richards GR. 2017. Stringent Response Regulators
864 Contribute to Recovery from Glucose Phosphate Stress in Escherichia coli.
865 Appl Environ Microbiol 83.
866 50. Vanderpool CK, Gottesman S. 2004. Involvement of a novel transcriptional
867 activator and small RNA in post-transcriptional regulation of the glucose
868 phosphoenolpyruvate phosphotransferase system. Mol Microbiol 54:1076-89.
869 51. Engels V, Wendisch VF. 2007. The DeoR-type regulator SugR represses
870 expression of ptsG in Corynebacterium glutamicum. J Bacteriol 189:2955-66.
871 52. Gaigalat L, Schlüter JP, Hartmann M, Mormann S, Tauch A, Pühler A,
872 Kalinowski J. 2007. The DeoR-type transcriptional regulator SugR acts as a
873 repressor for genes encoding the phosphoenolpyruvate:sugar
874 phosphotransferase system (PTS) in Corynebacterium glutamicum. BMC Mol
875 Biol 8:104.
876 53. Tanaka Y, Teramoto H, Inui M, Yukawa H. 2008. Regulation of expression of
877 general components of the phosphoenolpyruvate: carbohydrate
878 phosphotransferase system (PTS) by the global regulator SugR in
879 Corynebacterium glutamicum. Appl Microbiol Biotechnol 78:309-18. 37
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
880 54. Blombach B, Arndt A, Auchter M, Eikmanns BJ. 2009. L-valine production
881 during growth of pyruvate dehydrogenase complex-deficient Corynebacterium
882 glutamicum in the presence of ethanol or by inactivation of the transcriptional
883 regulator SugR. Appl Environ Microbiol 75:1197-200.
884 55. Repaske R, Clayton MA. 1978. Control of Escherichia coli growth by CO2. J
885 Bacteriol 135:1162-4.
886 56. Valley G, Rettger LF. 1927. The Influence of Carbon Dioxide on Bacteria. J
887 Bacteriol 14:101-37.
888 57. Bracher JM, Martinez-Rodriguez OA, Dekker WJC, Verhoeven MD, van Maris
889 AJA, Pronk JT. 2019. Reassessment of requirements for anaerobic xylose
890 fermentation by engineered, non-evolved Saccharomyces cerevisiae strains.
891 FEMS Yeast Res 19.
892 58. Gonzalez JE, Long CP, Antoniewicz MR. 2017. Comprehensive analysis of
893 glucose and xylose metabolism in Escherichia coli under aerobic and
894 anaerobic conditions by (13)C metabolic flux analysis. Metab Eng 39:9-18.
895 59. Kalinowski J, Bathe B, Bartels D, Bischoff N, Bott M, Burkovski A, Dusch N,
896 Eggeling L, Eikmanns BJ, Gaigalat L, Goesmann A, Hartmann M, Huthmacher
897 K, Krämer R, Linke B, McHardy AC, Meyer F, Möckel B, Pfefferle W, Pühler A,
898 Rey DA, Rückert C, Rupp O, Sahm H, Wendisch VF, Wiegräbe I, Tauch A.
899 2003. The complete Corynebacterium glutamicum ATCC 13032 genome
900 sequence and its impact on the production of L-aspartate-derived amino acids
901 and vitamins. J Biotechnol 104:5-25.
902 60. Keilhauer C, Eggeling L, Sahm H. 1993. Isoleucine synthesis in
903 Corynebacterium glutamicum: Molecular analysis of the ilvB-ilvN-ilvC operon.
904 J Bacteriol 175:5595-5603.
38
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
905 61. Kensy F, Zang E, Faulhammer C, Tan RK, Büchs J. 2009. Validation of a
906 high-throughput fermentation system based on online monitoring of biomass
907 and fluorescence in continuously shaken microtiter plates. Microbial Cell
908 Factories 8:1-17.
909 62. Mustafi N, Gr A, Kohlheyer D, Bott M, Frunzke J. 2012. The development and
910 application of a single-cell biosensor for the detection of L-methionine and
911 branched-chain amino acids. Metabolic Engineering 14:449-457.
912 63. Sambrook J, Russell DW. 2001. Molecular Cloning: A Laboratory Manual, 3
913 ed. Cold Spring Harbor Laboratory Press.
914 64. Eikmanns BJ, Thum-Schmitz N, Eggeling L, Lüdtke KU, Sahm H. 1994.
915 Nucleotide sequence, expression and transcriptional analysis of the
916 Corynebacterium glutamicum gltA gene encoding citrate synthase.
917 Microbiology 140:1817-1828.
918 65. Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, Smith HO. 2009.
919 Enzymatic assembly of DNA molecules up to several hundred kilobases.
920 Nature Methods 6:343-345.
921 66. Schäfer A, Tauch A, Jäger W, Kalinowski J, Thierbach G, Pühler A. 1994.
922 Small mobilizable multi-purpose cloning vectors derived from the Escherichia
923 coli plasmids pK18 and pK19: selection of defined deletions in the
924 chromosome of Corynebacterium glutamicum. Gene 145:69-73.
925 67. Van der Rest ME, Lange C, Molenaar D. 1999. A heat shock following
926 electroporation induces highly efficient transformation of Corynebacterium
927 glutamicum with xenogeneic plasmid DNA. Appl Microbiol Biotechnol 52:541-
928 545.
39
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
929 68. Niebisch A, Bott M. 2001. Molecular analysis of the cytochrome bc1-
930 aa3branch of the Corynebacterium glutamicum respiratory chain containing an
931 unusual diheme cytochrome c1. Arch Microbiol 175:282-294.
932 69. Kabus A, Georgi T, Wendisch VF, Bott M. 2007. Expression of the Escherichia
933 coli pntAB genes encoding a membrane-bound transhydrogenase in
934 Corynebacterium glutamicum improves L-lysine formation. Appl Microbiol
935 Biotechnol 75:47-53.
936 70. Baumgart M, Luder K, Grover S, Gätgens C, Besra GS, Frunzke J. 2013.
937 IpsA, a novel LacI-type regulator, is required for inositol-derived lipid formation
938 in Corynebacteria and Mycobacteria. BMC Biology 11.
939 71. Paczia N, Nilgen A, Lehmann T, Gätgens J, Wiechert W, Noack S. 2012.
940 Extensive exometabolome analysis reveals extended overflow metabolism in
941 various microorganisms. Microbial Cell Factories 11:1-14.
942 72. Hummel J, Strehmel N, Selbig J, Walther D, Kopka J. 2010. Decision tree
943 supported substructure prediction of metabolites from GC-MS profiles.
944 Metabolomics 6:322-333.
945 73. Kinoshita S, Udaka S, Shimono M. 2004. Studies on the amino acid
946 fermentation. Part 1. Production of L-glutamic acid by various microorganisms.
947 J Gen Appl Microbiol 50:331-343.
948 74. Schwentner A, Feith A, Münch E, Busche T, Rückert C, Kalinowski J, Takors
949 R, Blombach B. 2018. Metabolic engineering to guide evolution - Creating a
950 novel mode for L-valine production with Corynebacterium glutamicum. Metab
951 Eng 47:31-41.
952 75. Cremer J, Eggeling L, Sahm H. 1990. Cloning the dapA dapB cluster of the
953 lysine-secreting bacterium Corynebacterium glutamicum. Mol Gen Genet
954 220:478-480. 40
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
955
41
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
956 Tables
957 Table 1: Bacterial strains used in this study.
Strain Characteristics Reference C. glutamicum Wild type, biotin-auxotroph S. Kinoshita et al. (73) ATCC 13032 ∆aceE ATCC 13032 derivate with deletion of aceE M. E. Schreiner et al. (5) ∆aceE ∆pyc ATCC 13032 derivate with deletion of aceE This work and pyc ∆aceE ∆ptsG ATCC 13032 derivate with deletion of aceE This work and ptsG ∆aceE ∆pyc ∆ptsG ATCC 13032 derivate with deletion of aceE This work and pyc and ptsG ∆aceE ∆pgi ATCC 13032 derivate with deletion of aceE This work and pgi ∆aceE ∆pyc ∆pgi ATCC 13032 derivate with deletion of aceE This work and pyc and pgi ∆aceE ∆ppc ATCC 13032 derivate with deletion of aceE This work and ppc ∆aceE ∆pyc ∆ppc ATCC 13032 derivate with deletion of aceE, This work pyc and ppc ∆aceE ureD- ATCC 13032 derivate with deletion of aceE R. Mahr et al. (34) E188* and ureD-E188* (Glu188 to stop codon) ∆aceE ureD- ATCC 13032 derivate with deletion of aceE, This work E188* ∆pyc pyc and ureD-E188* (Glu188 to stop codon) ∆pyc ATCC 13032 derivate with deletion of pyc P. G. Peters- Wendisch et al. (24) ∆pyc ∆ppc ATCC 13032 derivate with deletion of aceE A. Schwentner et al. and ppc (74) E. coli DH5α fhuA2 lac(del)U169 phoA glnV44 Φ80' Invitrogen lacZ(del)M15 gyrA96 recA1 relA1 endA1 thi- 1 hsdR17; for general cloning purposes 958
959
42
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
960 Table 2: Plasmids used in this study.
Plasmid Characteristics Reference pK19mobsacB Contains negative (sacB) and positive (kanR) selection A. Schäfer marker for genomic integration and deletion et al. (66) pK19mobsacB- KanR; derivative of pK19mobsacB for partial pyc gene P. G. ∆pyc deletion, contains only the first 499 bp and last 493 bp of Peters- pyc, for the deletion of 2429 bp in the middle Wendisch et al. (24) pK19mobsacB- KanR; derivative of pK19mobsacB for partial ptsG gene This work ∆ptsG deletion, contains only the last 491 bp of ptsG, for the deletion of the first 1561 bp pK19mobsacB- KanR; derivative of pK19mobsacB for pgi gene deletion, This work ∆pgi contains only the last 537 bp of pgi, for the deletion of the
first 1086 bp pK19mobsacB- KanR; derivative of pK19mobsacB for partial ppc gene J. Buchholz ∆ppc deletion, contains only the first 391 bp and last 447 bp of et al. (13) ppc, for the deletion of 1922 bp in the middle pK19mobsacB- KanR; derivative of pK19mobsacB for ureD gene deletion, R. Mahr et ureD-E188* exchange of Glu188 to stop codon in ureD al. (34)
R pJC1 Kan , oriVEc, oriVCg; E. coli - C. glutamicum shuttle vector J. Cremer et al. (75)
pJC1-venus- KanR; derivative of pJC1, contains terminator sequence of M. term-BS Bacillus subtilis behind venus Baumgart et al. (70) pJC1-pyc KanR; derivative of pJC1-venus-term-BS, contains pyc This work under control of its native promoter Ppyc
R pJC1-lacI/Ptac- Kan ; derivative of pJC1-venus-term-BS, contains aceA This work aceA-linker- followed by a linking sequence (5’- aceB ACTAGAAATAATTTTGTTTAACTTTAAGAAGGAGATATACAT-
3’) and aceB under control of the inducible promoter Ptac R pJC1-lacI/Ptac- Kan ; derivative of pJC1-venus-term-BS, contains aceB This work aceB-linker- followed by a linking sequence (5’- aceA ACTAGAAATAATTTTGTTTAACTTTAAGAAGGAGATATACAT-
3’) and aceA under control of the inducible promoter Ptac 961
962
43
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
963 Table 3: Glucose consumption rate of C. glutamicum PDHC-deficient strains in the
964 exponential growth phase. For an overview of glucose consumption of the entire time
965 course of the experiment including the stationary phase, the reader is referred to Fig.
966 S2.
Strain Glucose consumption rate (nmol min-1 g-1)a ΔaceE 16.81 ± 3.01 ΔaceE Δpyc 10.87 ± 4.95 ΔaceE ΔptsG 7.24 ± 3.97 ΔaceE Δpyc ΔptsG 7.32 ± 1.88 967
968 a Means ± standard deviation of three independent biological replicates.
969
970
44
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
971 Table 4: Overview of lag phases and growth rates of C. glutamicum ΔaceE and
972 ΔaceE Δpyc cultivated on different TCA cycle carbon sources in CGXII medium
973 containing 154 mM acetate and 222 mM glucose.
ΔaceE ΔaceE Δpyc Added carbon source* Lag Growth rate (h- Lag Growth Δt (%)** Δt (%)** phase 1) phase rate (h-1) 0.27 ± / - 0 0.41 ± 0.01 + 100 0.02 0.33 ± Glu-Ala (3 mM) - 0 0.35 ± 0.01 ÷ 10 0.01 0.35 ± Citrate (50 mM) - 0 0.39 ± 0.01 ÷ 40 0.01 0.34 ± Succinate (100 mM) - -8 0.36 ± 0.01 ÷ 15 0.02 0.35 ± Malate (100 mM) - -8 0.35 ± 0.01 ÷ 38 0.02 974
975 *Additionally the medium contained 154 mM acetate and 222 mM glucose,
976 ** The difference between the first doubling time of strain ΔaceE and ΔaceE Δpyc
977 was defined as 100%. + = elongated lag phase referring to growth on 154 mM
978 acetate with 222 mM glucose, ÷ = shorter lag phase, - = no lag phase.
979
980
45
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
981 Table 5: Overview of lag phases and growth rates of C. glutamicum Δpyc, Δpyc Δppc
982 and ΔaceE Δpyc Δppc cultivated on different TCA cycle carbon sources in CGXII
983 medium containing 154 mM acetate and 222 mM glucose.
Δpyc Δpyc Δppc ΔaceE Δpyc Δppc Added carbon Growth rate Growth rate Growth rate source* Δt ** Δt ** Δt ** lag (h-1)*** lag (h-1)*** lag (h-1)*** / 6 h 0.17 ± 0.02 No growth No growth
Glu-Ala n.d. n.d. 8 h 0.21 ± 0.01 30 h 0.14 ± 0.01 (3 mM) Citrate 3 h 0.19 ± 0.01 5 h 0.26 ± 0.01 8 h 0.20 ± 0.03 (50 mM) Succinate 2 h 0.2 ± 0.02 3 h 0.16 ± 0.03 5 h 0.12 ± 0.03 (100 mM) Malate 3 h 0.21 ± 0.01 17 h 0.18 ± 0.01 18 h 0.24 ± 0.01 (100 mM) 984
985 *Additionally the medium contained 154 mM acetate and 222 mM glucose,
986 ** = Δtlag represents the difference in lag phase duration (h) between ΔaceE growing
987 on glucose-acetate mixture and the strain of interest growing on the respective
988 carbon source(s).
989 *** = for comparison, growth rate in glucose-acetate mixture was 0.41 h-1 for ΔaceE
-1 990 and 0.28 h for ΔaceE Δpyc in this experiment. n.d. = not determined
991
992
46
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
993 Table 6: Comparative transcriptome analysis of C. glutamicum ΔaceE Δpyc and
994 strain ΔaceE and during growth on glucose and acetate. Shown are selected genes
995 encoding for central metabolic enzymes or regulators (for mRNA ratios of ΔaceE
996 Δpyc versus ΔaceE ≤ 0.70 values are shaded in red; mRNA ratio ≥ 1.70 are shaded
997 in green). p-values >0.05 are highlighted red (n=3 biological replicates). For the
998 complete dataset, the reader is referred to Table S3.
Gene Gene Annotation mRNA p-Value name ratio PTS-components cg1537 ptsG glucose-specific enzyme II BC component of PTS 0.59 0.02 cg2117 ptsI PEP:sugar phosphotransferase system enzyme I 0.67 0.01 cg2121 ptsH phosphocarrier protein HPR of PTS 1.16 0.11
Myo-Inositol transporter
cg0223 iolT1 myo-Inositol transporter 1.97 0.04 cg3387 iolT2 myo-Inositol transporter 1.41 0.19
Transcriptional regulators of ptsG
cg2783 gntR1 transcriptional regulator, GntR family 0.76 0.05 cg1935 gntR2 transcriptional regulator, GntR family 0.97 0.46 cg2115 sugR DeoR-type transcriptional regulator of ptsG, ptsS and 0.79 0.03 ptsF
Acetate metabolism
cg3047 ackA acetate/propionate kinase 1.99 0.01 cg3048 pta phosphate acetyltransferase 1.34 0.04 cg2560 aceA isocitrate lyase 1.70 0.12
cg2559 aceB malate synthase 1.88 0.02
Anaplerosis
cg1787 ppc phosphoenolpyruvate carboxylase 0.91 0.32 cg3169 pck phosphoenolpyruvate carboxykinase (GTP) 0.72 0.03 cg1458 odx oxaloacetate decarboxylase 1.07 0.32 cg3335 malE malic enzyme 0.48 0.26
Further enzymes of central carbon metabolism
cg0973 pgi glucose-6-phosphate isomerase 0.44 0.03
47
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
cg2291 pyk pyruvate kinase 0.59 0.01 cg2891 pqo pyruvate chinon oxidoreductase 0.47 0.02 999
48
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1000 Table 7: Key mutations identified in adaptive laboratory evolution experiment of
1001 C. glutamicum ΔaceE Δpyc.
Single clone Batch No. Gene Gene Annotation Mutation ΔaceE Δpyc ID name 1 7 cg2117 ptsI EI enzyme, general component Exchange of PTS (EC: 2.7.3.9) R132H 2 8 cg0601 rpsC 30 S ribosomal protein S3 Exchange R225H 3 10 cg2117 ptsI EI enzyme, general component Exchange of PTS (EC: 2.7.3.9) G15S
1002
1003
49
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1004 Figure Legends
1005
1006 1007 Figure 1: Schematic overview of the central carbon metabolism of C.
1008 glutamicum. The main aspects of glycolysis, gluconeogenesis, pentose phosphate
1009 pathway, tricarboxylic acid cycle (TCA) and anaplerosis are shown, while grey boxes
1010 represent relevant enzymes. Carboxylation as well as decarboxylation steps are
1011 given. ABCRib = ATP-binding cassette transporter for ribose, AK = acetate kinase, CA
1012 = carbonic anhydrase, , GntP = gluconate permease, GPDH = glucose-6P
1013 dehydrogenase, ICD = Isocitrate dehydrogenase, KGDH = α-ketoglutarate
1014 dehydrogenase complex, PCx = pyruvate carboxylase, PDHC = pyruvate
1015 dehydrogenase complex, PEPCk = PEP carboxykinase, PEPCx = PEP carboxylase,
1016 PGDH = 6-P gluconate dehydrogenase, PGI = phosphoglucose isomerase, PK =
50
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1017 pyruvate kinase, PQO = pyruvate:quinone oxidoreductase, PTA =
1018 phosphotransacetylase, PTSGlc = permease of phosphotransferase system for
1019 glucose, PTSFru= permease of phosphotransferase system for fructose.
1020
51
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1021 1022 Figure 2: Impact of glucose on aceE and pyc mutant strains. Growth curves
1023 shown are based on the backscatter measurements in a microtiter cultivation system,
1024 inoculated at an OD600 of 1. Data represent the average of three biological replicates,
1025 error bars the standard deviation. (A) The C. glutamicum strains ∆aceE and ∆aceE
1026 ∆pyc, as well as the wild type were inoculated in CGXII media containing 154 mM
1027 acetate (left) or 154 mM acetate plus 222 mM glucose (right). (B) Strains ΔaceE and
1028 ΔaceE Δpyc as well as WT cells were cultivated in CGXII medium containing 154 mM
52
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1029 acetate and 222 mM glucose and supplemented either with (red lines) or without 100
- 1030 mM KHCO3 (blue and black lines). The box zooms in the time interval between 4 h
1031 and 15 h. (C) Strains ΔaceE Δpyc (left) and Δpyc (right) were transformed with pJC1-
1032 pyc plasmid for complementation of the pyc deletion or with pJC1-empty vector as
1033 control. Cultures were inoculated in CGXII containing 154 mM acetate, 222 mM
1034 glucose and 25 µg/ml kanamycin.
1035
53
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1036 1037 Figure 3: Influence of glucose consumption on the growth of C. glutamicum
1038 ΔaceE and ΔaceE Δpyc. Growth curves shown are based on the backscatter
1039 measurements in a microtiter cultivation system, inoculated at an OD600 of 1. (A) The
1040 strains ∆aceE (left) and ∆aceE ∆pyc (right) were inoculated to an OD600 of 1 in CGXII
1041 medium containing 154 mM acetate and either 0 mM (red), 100 mM (grey), 222 mM
1042 (blue) or 250 mM glucose (orange). (B) Deletion of the ptsG gene in the ΔaceE and
1043 ΔaceE Δpyc strains restored growth on glucose acetate mixtures (shades of red)
1044 when compared to the parental strains ΔaceE and ΔaceE Δpyc (shades of blue).
1045 Data represent the average of three biological replicates, error bars the standard
1046 deviation.
54
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1047 - 1048 Figure 4: Increased availability of HCO3 improved anaplerotic flux via PEPCx in
1049 pyc-deficient C. glutamicum strains. Growth curves shown are based on the
1050 backscatter measurements in a microtiter cultivation system, inoculated at an OD600
1051 of 1. Symbols represent the backscatter means of biological triplicates (n=3). (A) 55
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1052 Strain ΔaceE Δpyc Δppc was inoculated in CGXII medium containing 154 mM
- 1053 acetate and 222 mM glucose (shown in dark red) and, if indicated, 100 mM KHCO3
1054 was added (shown in light red). (B) The strains ΔaceE and ΔaceE Δpyc as well as
1055 the WT were cultivated in CGXII medium containing 154 mM acetate and 222 mM
1056 glucose adjusted to a pH of 6 (left) or pH of 8 (right) in comparison to cultivation at pH
1057 7. (C) The strains ΔaceE and ΔaceE ureD-E188* Δpyc were cultivated in CGXII
1058 media containing 154 mM acetate (left) or 154 mM acetate and 222 mM glucose
- 1059 (right) with and without the addition of KHCO3 . (D) The strains ΔaceE, ΔaceE Δpyc
1060 (left), ΔaceE ureD-E188* Δpyc (right) were inoculated in different growth media
1061 containing 154 mM acetate and 222 mM glucose (shades of blue), 100 mM gluconate
1062 (red) or 100 mM ribose (light red).
1063
56
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1064 - 1065 Figure 5: The intracellular CO2/HCO3 pool significantly affects L-valine
1066 production. Investigations on the influence of carbonate on growth and L-valine
1067 production of C. glutamicum ΔaceE and ΔaceE ureD-E188* are shown. L-valine
1068 (filled bars) and L-alanine (dashed bars) titers in the supernatant after 28 h of
1069 cultivation in the microtiter cultivation system. Data represent average values from
1070 three independent biological replicates using ultra-high performance liquid
1071 chromatography (uHPLC).
1072
57
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1073 1074 Figure 6: Refueling the TCA cycle to complement glucose intolerance of ΔaceE
1075 Δpyc. Growth curves shown are based on the backscatter measurements in a
1076 microtiter cultivation system, while symbols represent the backscatter means of
1077 biological triplicates (n=3). Strains ∆aceE and ∆aceE ∆pyc were inoculated to an
1078 OD600 of 1 in CGXII medium with 154 mM acetate and 222 mM glucose. (A) The TCA
1079 fueling carbon sources glutamate (3 mM Glu-Ala dipeptide), citrate (50 mM),
58
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1080 succinate (100 mM) or malate (100 mM) were added to the medium to analyze their
1081 effect on lag phase complementation (shades of red). As a control, growth on 154
1082 mM acetate and 222 mM glucose is shown for ΔaceE (dark blue) and ΔaceE Δpyc
1083 (light blue) in each experiment. (B) Strains ∆aceE and ∆aceE ∆pyc were transformed
1084 with pJC1-aceA-linker-aceB or pJC1-aceB-linker-aceA plasmid for simultaneous
1085 overexpression of the glyoxylate shunt enzymes isocitrate lyase (aceA) and malate
1086 synthase (aceB), while transformation of ΔaceE Δpyc and Δpyc with pJC1-empty
1087 vector served as control. Cultures were inoculated to an OD600 of 1 in CGXII
1088 containing 154 mM acetate, 222 mM glucose and 25 μg/ml kanamycin.
59
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1089 1090 Figure 7: Heterogeneous adaptation of C. glutamicum ΔaceE and ΔaceE Δpyc
1091 during growth on glucose and acetate. To identify a potential growing
1092 subpopulation the fluorescent dye PKH67 was used for membrane staining of ∆aceE
1093 and ∆aceE Δpyc prior to cultivation. Cells were cultivated in a microtiter cultivation
1094 system in the presence of 154 mM acetate and 222 mM glucose. As non-growing
1095 control, ∆aceE Δpyc was additionally incubated in CGXII containing solely 222 mM
1096 glucose. Cell growth (A) was online monitored using a microtiter cultivation system
60
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1097 and staining intensities were measured on single cell level during cultivation using
1098 flow cytometry (n=3 biological replicates). (B) Shown is the frequency of PKH67
1099 staining intensities distributions of one representative culture of each sample over the
1100 time course of cultivation (A).
1101
61
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1102 1103 Figure 8: Adaptive laboratory evolution of C. glutamicum ΔaceE Δpyc. Growth
1104 analysis of single inoculation steps obtained from the adaptive laboratory evolution
1105 (ALE) approach, which was performed with the strain ∆aceE ∆pyc in the presence of
1106 glucose and acetate. Growth curves are shown based on the backscatter
1107 measurements in a microtiter cultivation system. Symbols represent the backscatter
1108 means. (A) For growth analysis on the population level, glycerol stocks which were
1109 prepared during the ALE experiment, were used for the inoculation of a first pre-
1110 culture in BHI supplemented with 51 mM acetate. The second pre-culture in CGXII
1111 containing 154 mM acetate was then used for inoculation of the main culture in CGXII
1112 containing 154 mM acetate and 222 mM glucose (start OD600 of 1). Shown is the
1113 development of the relative lag phase duration after repetitive inoculations of ΔaceE
1114 Δpyc in media containing glucose and acetate. The difference between first doubling
1115 time of ΔaceE and ΔaceE Δpyc was set to 100%. For each inoculation, the difference
1116 of the lag phase compared to ΔaceE was calculated and is given in percent. (B)
1117 Single clones were isolated from batch number 7 (clone 1, black), 8 (clone 2, yellow)
1118 and 10 (clone 3, green) and were cultivated as mentioned above. ∆aceE (dark blue)
62
bioRxiv preprint doi: https://doi.org/10.1101/663856; this version posted June 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1119 and ∆aceE ∆pyc (light blue) served as control. These single clones were further
1120 analyzed by genome sequencing (Table 7).
63