The cellular response towards lanthanum is substrate specific and reveals a novel route for glycerol metabolism in Pseudomonas putida KT2440

Ever since the discovery of the first rare earth element (REE)-dependent enzyme, the physiological role of lanthanides has become an emerging field of research due to the potential environmental implications and biotechnological opportunities. In Pseudomonas putida KT2440, the two pyrroloquinoline quinone-dependent alcohol dehydrogenases (PQQ-ADHs) PedE and PedH are inversely produced in response to La3+-availability. This REE-switch is orchestrated by a complex regulatory network including the PedR2/PedS2 two-component system and is important for efficient growth on several alcoholic volatiles. As P. putida is exposed to a broad variety of organic compounds in its natural soil habitat, the cellular responses towards La3+ during growth on various carbon and energy sources were investigated with a differential proteomic approach. Apart from the Ca2+-dependent enzyme PedE, the differential abundance of most other identified proteins was conditional and revealed a substrate specificity. Concomitant with the proteomic changes, La3+ had a beneficial effect on lag-phases while causing reduced growth rates and lower optical densities in stationary phase during growth on glycerol. When these growth phenotypes were evaluated with mutant strains, a novel metabolic route for glycerol utilization was identified that seems to be functional in parallel with the main degradation pathway encoded by the glpFKRD operon. The newly discovered route is initiated by PedE and/or PedH, which most likely convert glycerol to glyceraldehyde. In the presence of lanthanum, glyceraldehyde seems to be further oxidized to glycerate, which, upon phosphorylation to glycerate-2-phosphate by the glycerate kinase GarK, is finally channelled into the central metabolism. Importance The biological role of rare earth elements has long been underestimated and research has mainly focused on methanotrophic bacteria. We have recently demonstrated that P. putida, a plant growth promoting bacterium that thrives in the rhizosphere of various feed crops, possesses a REE-dependent alcohol dehydrogenase (PedH), but knowledge about lanthanide-dependent effects on physiological traits in non-methylotrophic bacteria is still scarce. This study demonstrates that the cellular response of P. putida KT2440 towards La3+ is mostly substrate specific and that during growth on glycerol, La3+ has a severe effect on growth parameters. We provide compelling evidence that the observed physiological changes are linked to the catalytic activity of PedH and thereby identify a novel route for glycerol metabolism in this biotechnological relevant organism. Overall, these findings demonstrate that lanthanides can alter important physiological traits of non-methylotrophic bacteria, which might consequently influence their competitiveness during colonization of various environmental niches.

Introduction upon REE binding. Although its exact cellular role has yet to be established, it has been 91 speculated to play a role in Ln 3+ -uptake. Further, homologous genes have only been identified 92 in the genome of some other species of Methylobacteria and Bradyrhizobia. 93 In addition to their functional role as metal cofactor, several studies have recently investigated 94 fragments and BamHI-digested pJOE6261.2 were then joined together using one-step 150 isothermal assembly (45). 151 152

Strain constructions 153
Deletion mutant strains were constructed as previously described (46) were resuspended in 1 ml sample buffer (150 mM Tris-HCl pH 6.8; 2 % SDS; 20 mM 169 dithiothreitol) and heated for 5 min at 95°C with gentle shaking. Subsequently, samples were 170 centrifuged for 15 min at 21000 x g and 4°C, and the supernatants were stored in new reaction 171 tubes at -20 °C. In a next step, proteins were precipitated using chloroform-methanol (47) and To 25 µg protein in 60 µl Tris-buffered (50 mM, pH 8.5) urea (6 M), DTT was added to a final 177 concentration of 10 mM to guarantee reduction of cysteines. Samples were incubated for 30 178 min at 56 °C under shaking at 1000 rpm. Alkylation of cysteines was performed by adding 30 179 mM iodoacetamide and incubation for 45 min at room temperature in the dark. Alkylation was 180 stopped by adding 50 mM DTT and samples were incubated for another 10 min at RT. 500 ng 181 LysC protease (Roche) in 50 mM Tris buffer (pH 8.5) was added and samples were digested 182 overnight at 30 °C. Next, the urea in the reaction mixture was diluted to 2 M by adding the 183 appropriate amount of Tris buffer (50 mM, pH 8.5). 1 µg trypsin (Roche) in Tris buffer (50 mM, 184 pH 8.5) was added and digestion was continued for 4 hours at 37 °C. The digest was stopped 185 by addition of 3 µl 10% (v/v) trifluoroacetic acid (TFA). Next, peptide mixtures were 186 concentrated and desalted on C18 stage tips (49)

Mass spectrometry analysis 190
NanoLC-ESI-MS/MS experiments were performed on an EASY-nLC 1200 system (Thermo 191 Fisher Scientific) coupled to a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific) 192 using an EASY-Spray nanoelectrospray ion source (Thermo Fisher Scientific). Tryptic peptides 193 were directly injected to an EASY-Spray analytical column (2 μm, 100 Å PepMapRSLC C18, 194 25 cm × 75 μm, Thermo Fisher Scientific) operated at constant temperature of 35 °C. Peptides 195 were separated at a flow rate of 250 nL/min using a 240 min gradient with the following profile: 196 2% -10% solvent B in 100 min, 10% -22% solvent B in 80 min, 22% -45% solvent B in 55 197 min, 45% -95% solvent B in 5 min and isocratic at 90% solvent B for 15 min. Solvents used 198 were 0.5 % acetic acid (solvent A) and 0.5% acetic acid in acetonitrile/ H2O (80/20, v/v, solvent 199 B). The Q Exactive Plus was operated under the control of XCalibur 3.0.63 software. MS 200 spectra (m/z = 300-1600) were detected in the Orbitrap at a resolution of 70000 (m/z = 200) 201 using a maximum injection time (MIT) of 100 ms and an automatic gain control (AGC) value of 202 the 10 most abundant peptide precursors in the Orbitrap using high energy collision 205 dissociation (HCD) fragmentation at a resolution of 17500, a normalized collision energy of 27 206 and an intensity threshold of 1.3 x 10 5 . Only ions with charge states from +2 to +5 were selected 207 for fragmentation using an isolation width of 1.6 Da. For each MS/MS scan, the AGC was set 208 at 5 x 10 5 and the MIT was 100 ms. Fragmented precursor ions were dynamically excluded for 209 30 s within a 5 ppm mass window to avoid repeated fragmentation. 210 211

Protein quantification and data analysis 212
Raw files were imported into MaxQuant (51) version 1.6.0.1 for protein identification and label-213 free quantification (LFQ) of proteins. Protein identification in MaxQuant was performed using 214 the database search engine Andromeda (52). MS spectra and MS/MS spectra were searched 215 against P. putida KT2440 protein sequence database downloaded from UniProt (53). 216 Reversed sequences as decoy database and common contaminant sequences were added 217 automatically by MaxQuant. Mass tolerances of 4.5 ppm (parts per million) for MS spectra and 218 20 ppm for MS/MS spectra were used. Trypsin was specified as enzyme and two missed 219 cleavages were allowed. Carbamidomethylation of cysteines was set as a fixed modification 220 and protein N-terminal acetylation and oxidation were allowed as variable modifications. The 221 'match between runs' feature of MaxQuant was enabled with a match time window of one 222 minute and an alignment time window of 20 minutes. Peptide false discovery rate (FDR) and 223 protein FDR thresholds were set to 0.01. 224 Statistical analysis including t-tests and principal component analysis (PCA) were performed 225 using Perseus software version 1.6.0.2 (54). Matches to contaminant (e.g., keratins, trypsin) 226 and reverse databases identified by MaxQuant were excluded from further analysis. Proteins 227 were considered for LFQ (label free quantification) if they were identified by at least two 228 peptides. First, normalized LFQ values from MaxQuant were log2 transformed. Missing values 229 were replaced from normal distribution using a width of 0.2 and a downshift of 2.0. Statistical 230 differences between two sample groups were determined using an unpaired t-test and a p-231 value < 0.01 and a regulation factor > 2 (log2 fold-change > 1) were considered as significant change in protein abundance. The mass spectrometry proteomics data will be deposited to the 233 ProteomeXchange Consortium via the PRIDE (55) partner repository (submitted). 234 235

Purification and activity measurement of PQQ-ADHs PedE and PedH 236
To measure the activity of the two PQQ-ADHs PedE and PedH, the enzymes were expressed 237 in E. coli BL21(DE3) cells using plasmids pMW09 and pMW10, and purified by affinity 238 chromatography as described elsewhere (12). The activities with the four substrates 2-239 phenylethanol, citrate, glucose and glycerol were determined at a concentration of 10 mM 240 using a previously described colorimetric assay (12) with one minor modification. To represent 241 the growth conditions, 1 µM La 3+ instead of 1 µM Pr 3+ was used as metal cofactor for PedH. The authors would like to thank Prof. Bernhard Hauer for his continuous support. The authors 251 further declare that the research was conducted in the absence of any commercial or financial 252 relationships that could be construed as a potential conflict of interest. 253

255
We have recently demonstrated that the two PQQ-ADHs PedE and PedH are inversely 256 regulated dependent on the presence of rare earth elements (REEs) and that a complex 257 signalling network, which includes the activity of the PedR2/PedS2 two-component system, 258 orchestrates this regulation (12,38). To identify whether a global cellular response of P. putida 259 KT2440 towards REEs beyond the regulation of the PQQ-ADHs exists, we used a comparative 260 proteomic analysis during growth on four different carbon and energy sources, namely 2-261 phenylethanol, glycerol, glucose, and citrate. 262 263

Evaluation of proteomics data 264
Proteins were extracted from cells of P. putida by SDS to enable extraction of cytoplasmic as 265 well as transmembrane proteins followed by label free nano-LC-MS/MS quantification. In total, 266 2771 proteins with at least two unique peptides and an FDR ≤ 1% were identified and quantified 267 by our proteomics approach, corresponding to approximately 50% of the P. putida KT2440 268 proteome. Principal component analysis revealed high reproducibility for sample replicates and 269 distinct patterns for the different carbon sources (Fig. S1). The majority of proteins was 270 increased or decreased in abundance in response to the different carbon and energy sources. 271 In contrast, minor differences were observed in the presence or absence of La 3+ during growth 272 on the same carbon and energy source. Proteins that exhibited a 2-fold or higher change in 273 abundance between different growth conditions and a p-value ≤ 0.01 were considered as 274 differentially abundant. 275

Effect of lanthanum on protein abundance during growth with different substrates 277
According to the aforementioned criteria, 56 proteins were identified as differentially abundant 278 comparing growing cells of P. putida in the presence and absence of La 3+ with different carbon 279 sources ( Fig. 1, Table 3, Table S3-S5). In these studies, only the Ca 2+ -dependent PQQ-ADH 280 showed an increased abundance in response to La 3+ during growth on glucose, glycerol, and 283 2-phenylethanol, whereas an uncharacterized pentapeptide repeat containing protein 284 (PP_2673) that is directly upstream of pedE showed a decreased abundance during growth 285 on glycerol and 2-phenylethanol ( Fig. 1). The remaining 53 proteins were only identified under 286 one specific growth condition ( Table 3, Table S3-S5). 287 During growth on 2-phenylethanol and glycerol, a majority of the identified proteins was 288 increased in abundance (80% and 70%) in response to La 3+ ( Table 3, Table S3). For glucose 289 and citrate this was different, as most of the identified proteins were found to be less abundant 290 in response to La 3+ (36% and 40% during growth on glucose and citrate; Table S4 and Table  291 S5). Notably, the majority of the identified proteins were related to metabolism according to the 292 cluster of orthologous protein groups (COG) database (56). To test whether the observed 293 conditional proteomic response is linked to PedE and/or PedH activity, we determined the 294 corresponding enzyme activities with all four carbon and energy sources ( Table 1). Apart from 295 the already known substrate 2-phenylethanol, PedE and PedH also showed activity with 296 glycerol, whereas no activity could be detected with citrate or glucose. 297 298

Effect of lanthanum during growth on glycerol 299
From our proteomic-and biochemical data, we speculated that PedE and PedH activity could 300 play a beneficial role during glycerol metabolism of P. putida KT2440. As the degradation 301 pathway and growth characteristics of this organism have been recently characterized in great 302 detail (57, 58), we wanted to have a closer look on the effect of La 3+ during growth on this 303 specific carbon and energy source. In these experiments ( Fig. 2A, Table 2), we consistently 304 observed a shorter lag-phase (λ) of the cultures in response to La 3+ -availability (10.2 ± 0.2 h 305 vs. 17.3 ± 0.2). Additionally, the corresponding values of the specific growth rates (µmax, 0.201 306 ± 0.004 vs. 0.341 ± 0.010 h -1 ) and the maximal OD600 in stationary phase (OD600 max ; 0.680 ± 307 0.014 vs. 0.884 ± 0.004) of the cultures differed in the presence or absence of La 3+ , 308 respectively. As the purified PedH enzyme showed a 3-fold higher specific activity towards 309 glycerol compared to PedE in vitro (0.9 ± 0.1 U mg -1 vs. 0.3 ± 0.1 U mg -1 ; Table 1), we speculated that this increased glycerol conversion by PedH could be the underlying cause for 311 the observed differences in growth parameters. When subsequently a ΔpedE ΔpedH strain 312 was analysed for growth in the presence and absence of La 3+ (Fig. 2B, Table 2), no significant 313 differences in λ and µmax were observed for the ΔpedE ΔpedH strain in response to La 3+ while, 314 although less profound, small differences in OD600 max were still detected. Further, under both 315 conditions the double deletion strain showed a lag-phase that was undistinguishable from that 316 of the parental strain in the absence of La 3+ but dramatically longer than that of the parental 317 strain in the presence of La 3+ (17.5 ± 0.3 h in the absence of La 3+ and 17.2 ± 0.4 h in the 318 presence of 10 µM La 3+ ). Interestingly, the growth rates under both conditions (0.276 ± 0.008 319 h -1 and 0.291 ± 0.012 h -1 ) were significantly higher (p < 0.01) than those of the parental strain 320 in presence of La 3+ while still being significantly below (p < 0.05) those of the parental strain in 321 the absence of La 3+ . 322 These results implied that the two PQQ-ADHs can indeed be beneficial for growth on glycerol 323 and that a functionally active PedH enzyme is the underlying cause for the La 3+ -dependent 324 differences in lag-times and growth rates and to some extent also for differences in OD600 max 325 of KT2440 cultures. As PedE and PedH as well as the remaining proteins that were found to 326 be differentially abundant in response to La 3+ during growth on glycerol, are not part of the 327 described degradation pathway in P. putida KT2440 (57, 58), we hypothesized that an 328 additional metabolic route exists (Fig. 3). Based on our proteomic data, this route could be 329 initiated by the activity of PedH and the oxidation of glycerol to glycolaldehyde. In the next 330 steps glycolaldehyde could be oxidized to glycerate by PedH, the aldehyde dehydrogenase 331 AldB-II, or the aldehyde oxidase complex composed of proteins PP_3621 (IorA-II), PP_3622 332 and PP_3623 (AdhB). After phosphorylation by the glycerate kinase GarK, glycerate-2-333 phosphate could eventually enter the central metabolism. 334 If such a metabolic route exists, a ΔglpFKRD deletion strain should still be able to grow with 335 glycerol as sole source of carbon and energy, whereas a ΔpedE ΔpedH ΔglpFKRD deletion 336 mutant should not. To test this scenario, the corresponding strains were constructed and grew on glycerol independent of the GlpFKRD pathway, although growth was dramatically 339 impaired compared to the parental strain or strain ΔpedE ΔpedH. When PedE and PedH were 340 additionally deleted, no growth was observed even after a prolonged incubation time of 5 d. 341 This supported our hypothesis that a metabolic route for glycerol next to the GlpFKRD pathway 342 exists and that this route is initiated by PedE and PedH, most likely by the oxidation of glycerol 343 to glyceraldehyde (Fig. 3). Given that the route further proceeds via glycerate and glycerate-344 2-phosphate, different cellular concentrations of these metabolites would be expected during 345 growth on glycerol in a mutant that is not able to use the GlpFKRD pathway. Interestingly, a 346 companion study to this work employed a metabolome analysis using glycerol-growing cells 347 of P. putida KT2440 and a ΔglpK deletion strain, which can only use the proposed novel route 348 via PedE and PedH (59). In their experiments, the authors observed that the glycerate 349 concentration measured for the ΔglpK strain was dramatically increased compared to the wild 350 type, whereas concentrations of glyceraldehyde and glyceraldehyde-3-phosphate were in the 351 same range for both strains. These data suggested that glycerate is indeed an intermediate 352 during glpFKRD-independent growth, and that the activity of downstream proteins represent 353 the bottleneck of the metabolic route leading to the observed accumulation. As our proteomic 354 data indicated the involvement of the predicted glycerate kinase GarK, we constructed a ΔgarK 355 deletion strain and speculated that this strain should lack the ability to phosphorylate glycerate 356 and would hence be incapable of channelling glycerate-2-phosphate into the central 357 metabolism. We indeed observed no growth of a ΔgarK mutant on glycerate even after 358 incubation of up to 5 d, while strain ΔpedE ΔpedH ΔglpFKRD could grow and reached OD 600 max 359 within 72 h of incubation under the condition tested (Fig. 4B). 360 When grown on glycerol, no significant effect on growth rates and lag-times as well as only a 361 minor, but significant, negative effect on the OD600 max (p < 0.01) was observed for the ΔgarK 362 deletion in the absence of La 3+ . In contrast, the same deletion caused a dramatic growth 363 impairment in the presence of La 3+ (Fig. 5A, Table 2) and consequently the stationary phase deduced from these data. It is however obvious that the growth rate was far below the one of 366 the parental strain in the presence of La 3+ . 367 Notably, some of the most severely upregulated proteins in response to La 3+ are either related 368 to stress, namely the multidrug efflux pump MexEF and the alkylhydroperoxide reductase 369 subunits AhpC and AhpF, or are enzymes that play no obvious roles within the proposed 370 metabolic route such as CalA, a predicted coniferyl alcohol dehydrogenase, and the glycolate 371 oxidase GlcDEF. To investigate the potential influence of the latter two enzymes, we 372 constructed and analysed the corresponding ΔcalA and ΔglcDEF mutants and tested their 373 growth pattern with glycerol ( Fig. 5B; Table 2). The ΔglcDEF mutant showed a growth 374 behaviour similar to the parental strain. In contrast, the ΔcalA strain exhibited a significantly 375 increased growth rate (p < 0.01) and higher OD600 max (p < 0.01) than the parental strain in the 376 presence of La 3+ while showing no significant differences in OD600 max and maximal growth rate 377 and only minor differences (p < 0.05) in lag-times in the absence of La 3+ (Fig. 5C; Table 2). 378 In the present study, the cellular responses of P. putida KT2440 towards La 3+ -availability during 380 growth on several carbon and energy sources were investigated. The only protein that showed 381 a differential abundance independent of the substrate used for growth was the Ca 2+ -dependent 382 PQQ-ADH PedE. This result is in line with data from a previous study (38), which demonstrated 383 that the La 3+ -induced downregulation of pedE is dependent on the PedS2/PedR2 two-384 component system that, based on our current observation, seems to be functional under all 385 tested conditions. The other two proteins that showed differential abundance under more than 386 one culture condition (PedH, PP_2673) are both also part of the ped gene cluster. Notably, the 387 carbon and energy sources under which these proteins were identified either represent 388 substrates of PedE and PedH, or can be converted by an enzyme that depends on the same 389 PQQ-cofactor, namely the glucose dehydrogenase Gcd. The remaining 53 proteins that 390 showed differential abundance in response to La 3+ were identified only during growth on one 391 specific carbon and energy source, suggesting a conditional regulation. For glycerol, we 392 provide striking evidence that the increased activity of PedH compared to PedE is the primary 393 cause for the observed proteomic and physiological changes during growth. 394 Thus far, the degradation of glycerol was described to start by the uptake via GlpF, 395 phosphorylation by GlpK, and subsequent GlpD-catalysed oxidation of glycerol-3-phosphate 396 to dihydroxyacetone-3-phosphate (58). In a next step, dihydroxyacetone-3-phosphate is 397 interconverted to glyceraldehyde-3-phosphate and enters the central metabolism. This 398 pathway is negatively regulated by the transcriptional regulator GlpR, and the de-repression 399 of the glpFKRD operon is believed to depend on the intracellular concentration of glycerol-3-400 phosphate, which finally impacts the lag-phase of cultures (57). As such, it was interesting to 401 find that growth on glycerol in the presence of La 3+ led to a shorter lag phase and lower growth 402 rate of the parent strain, and that a ∆pedE ∆pedH mutant showed a lag-phase similar to the 403 parent in absence of La 3+ without any beneficial effect of La 3+ while still growing with a higher 404 growth rate than the parent strain in presence of La 3+ . Further experiments revealed that a not. Together with the notion that a ΔgarK mutant cannot utilize glycerate, this strongly 407 indicates the existence of a novel route for glycerol metabolism, in which PedE and PedH 408 catalyse the initial oxidation of glycerol to glyceraldehyde. In the presence of La 3+ , the route 409 seems to proceed via a second oxidation step to glycerate, which is subsequently converted 410 to glycerate-2-phosphate by the activity of GarK (Fig. 3). The PedE/PedH-dependent route, 411 despite being important for efficient growth, clearly is not the main route for glycerol 412 metabolism, as the effect of the ΔpedE ΔpedH deletion on the lag-phase with glycerol is far 413 less severe than deletion of the glpFKRD gene cluster. It also appears that the PedE/PedH-414 dependent route is less efficient than the GlpFKRD pathway, as the overall growth of the 415 ΔglpFKRD strain is substantially impaired in comparison to the ΔpedE ΔpedH strain. 416

A possible explanation could be the formation of the toxic intermediate glyceraldehyde, which 417
is known for its protein crosslinking properties and the formation of superoxide radicals due to 418 auto-oxidation (60, 61). The observed differences in growth rates and OD600 max in response to 419 La 3+ in the parent strain could thus reflect the increased metabolic flux towards glyceraldehyde 420 due to the higher specific activity of PedH compared to PedE. This would also explain the 421 severe La 3+ -dependent growth impairment of the ΔgarK mutant, as one can assume that even 422 higher concentrations of glyceraldehyde accumulate in a mutant that cannot process glycerate. 423 The notion that the MexEF RND-type transporter proteins, which are involved in efflux of 424 various toxic compounds (62), and the alkylhydroperoxide reductase subunits AhpC and AhpF, 425 which have been linked to ROS detoxification in P. putida (63), were also more abundant in 426 presence of La 3+ during growth on glycerol are supportive of such a hypothesis. 427 To explain the impact of La 3+ on the lag-times of cultures, one could speculate that in addition 428 to glycerol-3-phosphate, also other phosphorylated derivatives, such as glycerate-2-429 phosphate, are able to relieve the repression of glpFKRD by GlpR. However, as in the absence 430 of La 3+ the growth phenotype of the ΔgarK mutant is indistinguishable from that of the parental 431 strain, and since the growth rate of the parent strain in absence of La 3+ is still significantly 432 higher than the growth rate of strain ∆pedE ∆pedH, we postulate that yet another metabolic 433 route is present that contributes to growth without affecting the lag-phase. This second route could proceed via the phosphorylation of glyceraldehyde to glyceraldehyde-3-phosphate by 435 the activity of a so-far unknown kinase. Whether both alternative routes to the GlpFKRD 436 pathway are functional in parallel or whether the metabolic flux via glycerate is exclusively 437 induced in the presence of La 3+ is currently unknown and would need to be tested in future 438 studies. Similarly, the question why proteins that cannot be associated to the newly discovered 439 routes for glycerol are among the most differentially abundant proteins in response to La 3+ 440 remains to be elucidated. It is however worthwhile noting that CalA and GlcDEF are either 441 known (GlcDEF) or predicted (CalA) by the PROSITE software tool 442 (https://prosite.expasy.org/) (64) to be catalytically active on 2-hydroxy acids. As such, 443 potential activities towards pathway intermediates such as glycerate cannot be excluded at the 444

moment. 445
From our data, the La 3+ -dependent proteomic and physiological changes during growth on 446 glycerol can be explained by a shift in metabolic flux resulting from the differences in specific 447 catalytic activities between PedH and PedE. A similar metabolic-driven interpretation can also 448 be used to explain the proteomic differences during growth on other carbon and energy 449 sources that are known to be substrates for PedE and PedH such as 2-phenylethanol. 450 However, this logic fails to explain the differences observed during growth on glucose and 451 citrate, as they do not represent substrates for PedE and/or PedH. Despite the fact that we 452 currently do not know the underlying cause for the conditional proteomic changes under these 453

conditions, it indicates the presence of additional effects of REEs beside the interaction with 454
PedH and PedS2/PedR2. Such effects could include the inhibition of protein functions by 455 mismetallation (65, 66), changes in the physiology of the outer membrane (67), or so far 456 unknown REE-dependent enzymes and regulator proteins. The latter explanation is of 457 particular interest, as two recent studies provide strong evidence that specific importers that 458 can transport Ln 3+ into the cytoplasm of methylotrophic bacteria do exist (41,68). 459 Altogether, the current study demonstrates that the utilization of REEs can influence important 460 physiological traits of P. putida, which could be highly beneficial in competitive environmental food crops could hence be partially the result of increased competitiveness of plant growth 463 promoting organisms such as P. putida during root colonization. This hypothesis is further 464 supported by a recent study, which found that Pseudomonads predominantly thrive on root 465 exudates in vivo and are hence enriched in the rhizosphere of Arabidopsis thaliana (10). As 466 such, it will be interesting to see what future research will add to the currently emerging theme 467 of REEs being an important micronutrient for methylotrophic and non-methylotrophic

35.
Pang X, Li D, Peng A. 2002. Application of rare-earth elements in the agriculture of 574    PilA -3.08 3.88 722 Figure S1: PCA comparing the four different carbon sources with and without La 3+ .

767
Different carbon sources are indicated by squares (2-phenylethanol), circles (citrate), 768 diamonds (glucose) and stars (glycerol). Samples with 10 µM La 3+ or without La 3+ in the 769 medium are shown in blue and red, respectively. Biological replicates are indicated in the same 770 colour. Samples can be separated according to different carbon sources while treatment with 771 La 3+ only showed minor effects. 772