Metabolic engineering of bacteria for environmental applications: construction of Pseudomonas strains for biodegradation of 2-chlorotoluene

Journal of Biotechnology 85 (2001) 103– 113 www.elsevier.com/locate/jbiotec

Metabolic engineering of bacteria for environmental applications: construction of Pseudomonas strains for biodegradation of 2-chlorotoluene Marı´a-Amparo Haro, Vı´ctor de Lorenzo * Centro Nacional de Biotecnologı´a CSIC, Campus de Cantoblanco, 28049 Madrid, Spain Received 6 March 2000; received in revised form 29 May 2000; accepted 29 June 2000

Abstract In this article, we illustrate the challenges and bottlenecks in the metabolic engineering of bacteria destined for environmental bioremediation, by reporting current efforts to construct Pseudomonas strains genetically designed for degradation of the recalcitrant compound 2-chlorotoluene. The assembled pathway includes one catabolic segment encoding the toluene dioxygenase of the TOD system of Pseudomonas putida F1 (todC1C2BA), which affords the bioconversion of 2-chlorotoluene into 2-chlorobenzaldehyde by virtue of its residual methyl-monooxygenase activity on o-substituted substrates. A second catabolic segment encoded the entire upper TOL pathway from pWW0 plasmid of P. putida mt-2. The enzymes, benzyl alcohol dehydrogenase (encoded by xylB) and benzaldehyde dehydrogenase (xylC) of this segment accept o-chloro-substituted substrates all the way down to 2-chlorobenzoate. These TOL and TOD segments were assembled in separate mini-Tn5 transposon vectors, such that expression of the encoded genes was dependent on the toluene-responsive Pu promoter of the TOL plasmid and the cognate XylR regulator. Such gene cassettes (mini-Tn5 [UPP2] and mini-Tn5 [TOD2]) were inserted in the chromosome of the 2-chlorobenzoate degraders Pseudomonas aeruginosa PA142 and P. aeruginosa JB2. GC-MS analysis of the metabolic intermediates present in the culture media of the resulting strains verified that these possessed, not only the genetic information, but also the functional ability to mineralise 2-chlorotoluene. However, although these strains did convert the substrate into 2-chlorobenzoate, they failed to grow on 2-chlorotoluene as the only carbon source. These results pinpoint the rate of the metabolic fluxes, the non-productive spill of side-metabolites and the physiological control of degradative pathways as the real bottlenecks for degradation of certain pollutants, rather than the theoretical enzymatic and genetic fitness of the recombinant bacteria to the process. Choices to address this general problem are discussed. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Metabolic engineering; TOL; TOD; Catabolic segment; 2-Chlorotoluene

1. Introduction: ventures and limitations of metabolic engineering * Corresponding author. Tel.: +34-91-5854536; fax: + 3491-5854506. E-mail address: [email protected] (V. de Lorenzo).

It is now 25 years since the European Federation of Biotechnology established a Working

0168-1656/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 5 6 ( 0 0 ) 0 0 3 6 7 - 9

104

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

Party to identify key factors that limit the usefulness of micro-organisms in biotechnology. Microbial physiology is at the interface between biological discovery and genetic engineering on the one hand and environmental biotechnology, biochemical engineering and the exploitation of microbial productivity on the other. The development of novel organisms for use in biodegradation has been a key challenge throughout this period. In particular, the growing understanding of the genetic and biochemical bases of the metabolism of xenobiotic compounds holds considerable potential for the development of recombinant microorganisms useful for the bioremediation of environmental pollution caused by urban and industrial activities (van der Meer et al., 1992; Timmis et al., 1994; Knackmuss, 1996; Reineke, 1998). One interesting case that portrays the ventures and limitations of the metabolic engineering of bacteria destined for environmental bioremediation is the case of chlorinated aromatics and, among these, chlorinated toluenes. Even the simpler substituted congeners of these chemical species are quite recalcitrant to natural biodegradation under aerobic conditions. Although no natural strains have been characterised so far which are able to degrade mono-substituted chlorotoluenes, the observed relaxed substrate range of some of the enzymes of the upper TOL pathway of the plasmid pWW0 of P. putida mt-2 for bioconversion of toluene into benzoate opens interesting possibilities for metabolic engineering (Vandenbergh et al., 1981; Haigler and Spain, 1989; Shaw and Harayama, 1990). The introduction of the TOL upper genes into the 3-chlorobenzoate degrader Pseudomonas sp. B13 (Abril et al., 1989; Brinkmann and Reineke, 1992), generated strains claimed to degrade 3- and 4-chlorotoluene and 3-chlorotoluene, respectively. In no case, however, was it possible to produce strains active on 2-chlorotoluene. The seminal work by Brinkmann and Reineke (1992) clearly showed that that the xylene monooxygenase complex encoded by the xylAM genes of the TOL plasmid was able to transform 3-chloro and 4-chlorotoluene into the corresponding chlorobenzylaldehydes. These could then be further converted into

chlorobenzoates by the action of XylB and XylC and channelled through the modified ortho-cleavage pathway of Pseudomonas sp. B13. This was not true, however, for 2-chlorotoluene, which was hardly recognised at all as a substrate by the TOL xylene monooxygenase. Interestingly, the same authors (Brinkmann and Reineke, 1992) observed that 2-chlorobenzylalcohol, the derivative of 2chlorotoluene after monooxydation of the methyl group of this compound, could be channelled through the broad substrate range TOL dehydrogenases encoded by xylB and xylC (Fig. 1) all the way down to 2-chlorobenzoate. Thus, the key activity that would enable an upper pathway for degradation of 2-chlorotoluene appeared to be one for oxidation of its methyl group into the corresponding alcohol. In 1992, it was reported that the toluene dioxygenase of the TOD system of P. putida F1 (todC1C2BA) had a residual methyl-monooxygenase activity on o-nitrotoluene, capable of conversion of this substrate to o-nitrobenzylalcohol (Robertson et al., 1992). More recently, Lehning et al. (1997) reported the biotransformation of 2-chlorotoluene into 2-chlorobenzylalcohol through the ring-dioxygenase activities of the P. putida F1 and the chlorobenzene degrader Burkholderia sp. PS12. These two microorganisms successfully attacked 2-chlorotoluene, yielding the corresponding benzylalcohol at very high efficiency (\ 70%) although further metabolism occurred at a low rate. In either case, the dioxygenases that catalyse the first step of toluene/benzene biodegradation appear to act as monooxygenases when the ortho position of the substrate is occupied by an electronegative substitution, thereby converting 2-chlorotoluene into 2-chlorobenzylalcohol. These observations provided the basis for engineering strains able to entirely degrade the otherwise recalcitrant 2chlorotoluene. In this context, we describe below ongoing efforts towards the construction of recombinant strains for metabolising 2-chlorotoluene. To this end, we have combined two catabolic segments recruited from the TOL and the TOD pathways of P. putida, resulting in a hybrid upper pathway

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

for bioconversion of 2-chlorotoluene into 2chlorobenzoate. The corresponding genetic determinants were then transferred onto the chromosome of two different 2-chlorobenzoate degrader strains, expected to provide a suitable lower pathway for complete mineralisation of the substrate. As shown herein, the resulting recombinant bacteria did have all the necessary genetic information and the required enzymatic activities on 2-chlorotoluene and its downstream intermediates. The analysis of the metabolites accumulated in the medium under various growth conditions pinpointed, not only kinetic and enzymatic bottlenecks which restricted the performance of the hybrid routes, but also barriers related to the energy balance of the process, the spill of side-products and the regulation of the degradative pathways. The conclusions of this work may apply to many other examples where genetic engineering has been employed to address otherwise intractable

105

cases of biodegradation of chloroaromatic compounds.

2. Rationale for a hybrid pathway for metabolism of 2-chlorotoluene The strategy of choice to produce Pseudomonas strains able to degrade 2-chlorotoluene is summarised in Fig. 1. The most salient feature of the scheme is the assembly of a hybrid upper pathway able to convert 2-chlorotoluene into 2-chlorobenzoate. This involves the mono-oxydation of the methyl group of 2-chlorotoluene by the broad substrate range toluene dioxygenase from P. putida F1, encoded by todC1C2BA genes. The resulting 2-chlorobenzylalcohol is then expected to be further oxydised to the corresponding aldehyde and acid through the action of the TOL benzylalcohol dehydrogenase (encoded by xylB) and benzalde-

Fig. 1. Assembly of a novel pathway for biodegradation of 2-chlorotoluene by engineered Pseudomonas strains. In a first stage, the residual methyl-monooxygenase activity of the toluene dioxygenase (TOD) enzyme of P. putida F1 on ortho-substituted toluene (encoded by the todC1C2BA genes) is employed to convert 2-chlorotoluene into 2-chlorobenzylalcohol. This product is then channeled into the TOL upper pathway, whose xylB and xylC genes (encoding respectively, benzylalcohol dehydrogenase and benzaldehyde dehydrogenase) convert it all the way to 2-chlorobenzoate. Finally, the P. aeruginosa strains that host the engineered pathway have an intrinsic ability to consume 2-chlorobenzoate. P. aeruginosa JB2 (Hickey and Focht, 1990) seems to degrade this compound through a pathway that involves production of a 3-chlorocatechol intermediate that can be funnelled towards a modified ortho pathway. On the contrary, strain P. aeruginosa PA142 makes a 1,2-dioxygenation that results in formation of plain catechol. This can be entered as such in the housekeeping ortho pathway (Romanov and Hausinger, 1994).

106

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

from Pu when bacteria are exposed, not only to pathway substrates such a toluene or xylenes, but

Fig. 2. (Continued)

hyde dehydrogenase (the product of xylC), respectively, from the upper pathway for m-xylene and toluene degradation of P. putida mt-2 (Shaw and Harayama, 1990). The predicted product of this hybrid pathway is 2-chlorobenzoate, which can be used as growth substrate by either of the two P. aeruginosa strains expected to provide a lower pathway. In one case (P. aeruginosa JB2), the pathway for 2-chlorobenzoate biodegradadation involves formation of 3-chlorocatechol, which appears to be channeled through a modified ortho cleavage route (Hickey and Focht, 1990). Alternatively, the P. aeruginosa PA142 strain expresses a benzoate dioxygenase able to attack positions 1 and 2 of 2-chlorobenzoate, such that the resulting hydrodiol is converted to catechol, releasing CO2 and chloride (Romanov and Hausinger, 1994). Besides providing all of the enzymes required for the degradation of 2-chlorotoluene, we also engineered the corresponding catabolic segments for coordinated expression of the TOL and TOD genes under the single control of the Pu promoter and the cognate XylR protein of the TOL plasmid (Fig. 3). XylR is able to activate transcription

Fig. 2. Design of catabolic segments encoding TOL and TOD activities. In the upper part of the figure, the organization of the xyl genes of the upper TOL operon is shown lined up with the restriction map of the portion of the pWW0 plasmid involved (Williams et al., 1997). Genes xylUW and xylN are not required for degradation of toluene but they form part of the same transcriptional unit. The location and orientation of the tolueneresponsive promoter of the system, Pu, is indicated also. The whole operon and its native promoter was engineered as a single NotI DNA segment from pCK04 (Panke et al., 1998) and inserted in plasmid pJMT6 (Sa´nchez-Romero et al., 1998), which places the upper pathway with a Telr mini-Tn5 transposon, whose organisation is sketched (Telr gene not to scale, 3 kb) flanked by the I and O ends of the mobile element. This element spanning the entire upper TOL pathway (Harayama et al., 1989; Williams et al., 1997; Panke et al., 1998) was constructed as follows. The 9.1 kb NotI insert of plasmid pCK04 (Panke et al., 1998) was cloned in vector pJMT6 (Sa´nchezRomero et al., 1998). The resulting plasmid (pUPP2) bears the whole xylUWCMABN operon preceeded by its native Pu promoter and all assembled within the boundaries of a Telr mini-Tn5 vector (mini-Tn5 [UPP2]). Restriction sites: N, NotI; E, EcoRI; Xm, XmaI; Xb, XbaI; H, HindIII; C, ClaI; A, AccI; Av, A6rII. Similarly, the organisation of the DNA segment that bears the todC1C2BA cluster encoding the multicomponent toluene dioxygenase of the TOD system (along with the hydrolase, todD) is shown in the lower part of the figure. The cluster was assembled as a NotI fragment and inserted in a vector that places its expression of the tod genes under the control of the Pu promoter and its cognate toluene-responsive regulator XylR, all assembled within the boundaries of a mini-Tn5 Sm/Sp mini-transposon (vector part of the scheme not to scale). To generate this construct, the 1576 bp EcoRI-HindIII fragment from plasmid pDTG602 (Zylstra et al., 1988), containing the whole todC1 gene and part of todC2, was cloned into the corresponding sites of the low copy number vector pCK01 (Ferna´ndez et al., 1995), originating plasmid pAV1. A separate 3135 bp HindIII fragment from pDTG602 (Zylstra and Gibson, 1989) which spanned the rest of todC2, as well as todB, todA, and todD, was cloned in the single HindIII site of pAV1 to reassemble the entire tod operon (todC1C2BAD), as a promoterless gene cluster flanked by NotI restriction sites. The resulting plasmid (pTOD1) did express a substantial dioxygenase activity, as noted by the dark (indigo) colour of E. coli CC118 (pTOD1) colonies grown on LB agar plates. To transfer the reconstructed tod operon to a suitable expression and integration system, a 4.7 kb NotI fragment generated from partial digestion of pTOD1 (an internal NotI site exists within the todA sequence) was cloned into vector pCNB3 (de Lorenzo et al., 1993). The resulting plasmid (pTOD2) bore the operon led by todC1 under the control of the toluene-responsive Pu promoter of the TOL plasmid. This ensures expression of the tod genes upon exposure of the host cells to this aromatic compound and its structural analogues (Brinkmann and Reineke, 1992; Abril et al., 1989). pTOD2 thus includes the entire set of genes todC1C2BAD, encoding the toluene dioxygenase complex and the cis-dihydrodiol dehydrogenase activity (Zylstra et al., 1988) within a mini-Tn5 mobile element (mini-Tn5 [TOD2]).

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

107

3. Construction and gross characterisation of P. aeruginosa strains bearing TOL and TOD genes

Fig. 3. Regulatory circuit for concurrent induction of xyl and tod genes with 2-chlorotoluene. The XylR protein is expressed from the chromosomally inserted mini-Tn5 [TOD2] element, whose organisation is outlined on top of the figure. Upon exposure to the substrate (2-chlorotoluene), XylR activates simultaneously both the adjacent tod genes as well as the Pu promoter upstream of the xyl genes present in the second mobile element, mini-Tn5 [UPP2], inserted in the same host P. aeruginosa strain.

also to a variety of structural analogues including 2-chlorotoluene (Abril et al., 1989; Brinkmann and Reineke, 1992). In contrast, the TOD genes of P. putida F1 are subjected to a two-component regulatory switch whose primary effector remains unknown (Lau et al., 1997). To overcome this uncertainty, we preferred to subject transcription of both the tod and the xyl gene clusters to the single control of XylR and 2-chlorotoluene, as shown in Fig. 3. This was achieved by placing the promoterless todC1C2BA gene cluster downstream of the Pu promoter of pCNB3, which also bears the xylR gene within the boundaries of the mobile element (mini-Tn5 [TOD2], Fig. 2). The other catabolic segment with the upper TOL genes present in mini-Tn5 [UPP2] inherited its own Pu promoter, regulated by the XylR protein provided in trans from mini-Tn5 [TOD2] (Fig. 3). Not much is known about the regulation of 2chlorobenzoate biodegradation in P. aeruginosa strains JB2 (Hickey and Focht, 1990) and PA142 (Romanov and Hausinger, 1994), but we assumed that the cognate promoters would respond adequately to the presence of this substrate in the medium.

The metabolic engineering strategies described above were used to construct strains AH001 and AH002 as follows. The transposable elements mini-Tn5 [TOD2] and mini-Tn5 [UPP2] are shown in Fig. 2. Plasmid pUPP2 (Table 1) is the delivery plasmid for hybrid transposon mini-Tn5 [UPP2], which bears the upper TOL genes driven by the Pu promoter adjacent to a selectable tellurite resistance marker. pTOD2 bears mini-Tn5 [TOD2], in which the todC1C2BA genes expressed from an upstream Pu promoter in the vector and adjacent to the xylR gene. These mobile elements were targeted towards the chromosome of 2chlorobenzoate degrading strains P. aeruginosa PA142 and P. aeruginosa JB2 through separate and successive triparental matings of the recipient strains with E. coli CC118lpir (pTOD2) and E. coli CC118lpir (pUPP2), respectively, along with the helper strain E. coli HB101 (RK600), as described in detail in de Lorenzo and Timmis (1994) and Panke et al. (1998). Exconjugants bearing the desired insertions were selected on agar plates prepared with the mineral medium described in the references above and 2-chlorobenzoate as the only carbon source. In addition, the plates contained Sm/Sp (streptomycin/spectinomycin) for selection of mini-Tn5 [TOD2] insertions and/or Tel (potassium tellurite) for selection of mini-Tn5 [UPP2]. The colonies with the expected phenotype (growth on 2-chlorobenzoate, Telr, Sm/Spr) were then examined for sensitivity to Pip (piperacilline) in order to distinguish authentic transposition of the mobile elements (Pips) from the cointegration of the whole transposon-delivery plasmid (Pipr). The strain P. aeruginosa PA142, bearing both mini-Tn5 [TOD2] and mini-Tn5 [UPP2], was designed P. aeruginosa AH001, whereas the equivalent insertions in P. aeruginosa JB2 gave rise to strain P. aeruginosa AH002 (Table 1). The transposition frequency of each mobile element was estimated to be 10 − 6 –10 − 7 for both mini-Tn5 transposons (i.e. two orders of magnitude lower than the frequency in the reference strain, P. putida KT2442; de Lorenzo and Timmis, 1994). Their order of insertion affected neither this figure

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

108

nor the final properties of the double-inserted strains. Since our attempts to directly select clones growing on 2-chlorotoluene as the only C source were not successful (see below), we resorted to plates with 2-chlorobenzoate amended with the selection agents (streptomycin, spectinomycin or potassium tellurite), so at least the performance of the lower pathway of the predicted route was always ensured. After all insertions had been ver-

ified, one clone derived from each parental strain was selected for further studies. As stated above, these were named P. aeruginosa AH001 (derived from P. aeruginosa PA142) and P. aeruginosa AH002 (from P. aeruginosa JB2). Neither of these two strains was able to grow in plates or in liquid medium with 2-chlorotoluene as the sole C source. However, they did contain functional catabolic segments which, separately, allowed

Table 1 Bacteria and plasmidsa Species and strains

Relevant genotype/phenotype

Escherichia coli K12 strains CC118 D(ara-leu), araD, DlacX74, galE, galK, phoA20, thi-1, rspE, rpoB, argE (Am), recA1 CC118 lpir

Nalr, Rifr CC118, lysogenized with lpir phage;

S17-1 lpir

F-, recA, hsdRM

+

, thi lpir phage lysogen RP4-2, (Tc::Mu, Km::Tn7)

Pseudomonas aeruginosa strains PA142 2-chlorobenzoate degrader

JB2

2-chlorobenzoate degrader

AH001

Spr Telr P. aeruginosa PA142 inserted with mini-Tn5 [TOD2] from pTOD2 and mini-Tn5 [UPP2] from pUPP2. 2-chlorotoluene co-metabolizer. Spr Telr. P. aeruginosa JB2 inserted with mini-Tn5 [TOD2] from pTOD2 and mini-Tn5 [UPP2] from pUPP2. 2-chlorotoluene co-metabolizer.

AH002 Plasmids pAV1 pCK01 pCK04

Cmr, pSC101 replicon, pCK01 derivative with a 1.6 kb EcoRI-HindIII insert containing todC1 and part of todC2 from pDTG02 Cmr pSC101 repicon, low copy number, 3.6 kb. MCS of pUC18 flanked by NotI sites, a-lac fragment pCK01 carrying Pu /xylCMABN genes as a NotI fragment.

pDTG602

Apr Sm/Spr oriV R6K, mob RP4 ori T, suicide delivery plasmid of transposon vector mini-Tn5 xylR/Pu Apr plasmid containing todC1C2BAD genes

pJMT6

Apr Telr suicide delivery plasmid of transposon vector mini-Tn5 Tel

pTOD1

Cmr, pSC101 replicon, pAV1 derivative carrying todC1C2BAD genes as NotI flanked fragment Apr Sm/Spr pCNB3 derivative with a NotI flanked insert carrying the promoterless tod C1C2BAD genes adjacent to xylR /Pu Apr Telr pJMT6 derivative with a NotI flanked insert containing Pu /XylCMABN genes

pCNB3

pTOD2 pUPP2

Reference/origin

(Manoil and Beckwith, 1985) (de Lorenzo et al., 1990) (de Lorenzo and Timmis, 1994) (Romanov and Hausinger, 1994) (Hickey and Focht, 1990) This study This study

This study (Ferna´ndez et al., 1995) (Panke et al., 1998) (de Lorenzo et al., 1993) (Zylstra et al., 1988) (Sa´nchez-Romero et al., 1998) This study This study This study

a Abbreviations: Nal: nalidixic acid, Sm: streptomicin, Kan: kanamicin, Tc: tetracyclin, Rif: rifampicine, Tp: trimethoprim, Tel: tellurite.

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

growth on toluene vapours and originated dark colonies when streaked onto in LB plates, thus suggesting that the newly introduced TOL and TOD genes were active. In spite of these observations, the lack of growth with the chlorinated substrate could mean a partial or total lack of activity of the three individual metabolic blocks of Fig. 1 or a failure to couple productively their performance.

4. Metabolism of 2-chlorotoluene by citrate-fed P. aeruginosa AH001 and P. aeruginosa AH002 In view of the results above, we set out to examine the performance of the hybrid pathway for degradation of 2-chlorotoluene under growth conditions that did not require its utilisation as sole growth substrate. Because XylR regulated the expression of the tod and xyl genes, the maximum expression was expected at stationary phase of growth (Cases et al., 1996). Thus, we cultured separately P. aeruginosa AH001 and P. aeruginosa AH002 in a minimal medium containing 0.2% citrate and let it grow for 2 days until early stationary phase. During that time, the absorbance (A600) of the culture changed from 0.05 to 0.5 (i.e. :10-fold), after which the growth rate slowed down still further. The combination of growth in a non-repressive carbon source and the onset of the stationary phase enabled the Pu promoter (which drives expression of the xyl and tod catabolic segments, Fig. 3) to reach its maximum activity (Cases et al., 1996). After this treatment, the cultures were exposed during 4 days to saturating vapours of 2-chlorotoluene. During that time, cultures experienced only a very minor increase in its A600 (presumably accompanied by little or no growth). Following incubation under these conditions, the bulk of organic products were extracted from the supernatants acidified and submitted to CG-MS analysis. Control cultures included non recombinant, parental strains P. aeruginosa PA142 and P. aeruginosa JB2, as well as samples not exposed to 2-chlorotoluene vapours. This analysis permitted the monitoring of the performance of the

109

new pathway incorporated in AH001 and AH002 strains, regardless of the growth on the recalcitrant substrate. Representative CG-MS results are shown in Fig. 4(A,B). Since the two Pseudomonas strains that hosted the engineered catabolic segments had an intrinsic ability to degrade 2-chlorobenzoate (which was maintained in both P. aeruginosa AH001 and P. aeruginosa AH002), our efforts were concentrated in examining the presence of intermediates predicted from the combined action of the tod and xyl genes (Fig. 1). The main reference was thus the retention time and the MS breakage pattern of the 2-chlorobenzoate standard, as shown in Fig. 4. The first noticeable feature of the analyses performed under the various conditions was the similar pattern of peaks yielded in GC by samples taken from P. aeruginosa AH001 and P. aeruginosa AH002, in spite of the known divergences in their way to degrade 2-chlorobenzoate. This suggested that the type of metabolites found in the cultures depended mostly on the activity of the engineered upper pathway. A search was made to match the MS fingerprints of each GC peak with predicted intermediates of the hybrid upper pathway. As shown in Fig. 4A, samples collected from P. aeruginosa AH001 cultures after exposure of citrate-grown cultures to 2-chlorotoluene unequivocally contained mid-mM (: 20–200 mM) concentrations of 2-chlorobenzoate as the major product present in the culture medium. While this result suggested the correct performance of the engineered pathway, it also indicated that the transient accumulation of 2-chlorobenzoate could be insufficient to support significant growth. This notion was further strengthened by the examination of the GC-MS peaks resulting from samples taken from cultures pre-grown previously in the presence of 2-chlorotoluene. In such a case (Fig. 4B), under these conditions further intermediates were detected including, not only 2-chlorobenzoate, but a variety of additional compounds. Some of these could be totally or partially identified. For instance, the peak numbered 1 in Fig. 4(B) presented an MS spectrum likely to correspond (\ 79% probabil-

110

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

Fig. 4.

ity) to 2-chlorobenzylalcohol, one of the expected intermediates of the engineered upper pathway. Other compounds could not be fully identified, but involved certainly chlorohydroxyl aromatic rings. This is because they contained the typical chlorine mass spectrum signature and their silylated forms could be detected in GC-MS after BSTFA [o-bis-(trimethylsilyl)trifluoroacetamide] derivatisation, which requires a pre-existing hydroxyl group. These chlorohydroxyl aromatic compounds were certainly produced by the activity of the products of the genes engineered in the catabolic segments, since none of the control nonrecombinant strains, nor P. aeruginosa AH001 and P. aeruginosa AH002 not exposed to 2chlorotoluene released them to their corresponding supernatants. In fact, while the presence of 2-chlorobenzoate and 2-chlorobenzylalcohol verified the performance of the upper route as shown in Fig. 1, the range of other chlorohydroxyl compounds indicated as well an important promiscuous ring-hydroxylation activity in the engineered strains leading to non-specific dioxygenation of various aromatic intermediates.

5. Conclusion: genetic, enzymatic and physiological bottlenecks in degradation of chloro-aromatic compounds In this work we have reported the performance of two Pseudomonas strains equipped with a set of genes and enzymatic activities which allow them to metabolise 2-chlorotoluene, along with a suitable regulatory circuit. Although the strains were separately able to convert 2-chlorotoluene via 2chlorobenzylalcohol to 2-chlorobenzoate and to grow as well on 2-chlorobenzoate as the only carbon source, they failed to use the initial substrate as the only C and energy source. On the contrary, when cells were grown in an alternative carbon source such as citrate, they had a measurable ability to convert 2-chlorotoluene to 2chlorobenzoate, thus confirming the performance of the hybrid upper pathway. Such a performance was, however, insufficient to support growth. This cannot be attributed to defects in the lower pathway present in both P. aeruginosa AH001 and P. aeruginosa AH002 since they both retained the ability of the parental strains to grow on 2-

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

chlorobenzoate. For the strains presented here, we employed enzymatic activities with good conversion rates of 2-chlorotoluene into 2-chlorobenzylalcohol (the todC1C2BAC products) and of 2-chlorobenzylalcohol into 2-chlorobenzoate (the xylB/C products). However, they failed to couple productively their substrate conversion levels for funnelling enough 2-chlorobenzoate into the lower pathway. Besides, part of the enzymatic potential of the hybrid upper pathway could have been spoiled by the presence in the medium of aromatic intermediates that could act as unspecific substrates of the TOD mono/dioxygenase activity (Fig. 4B), leading to dead-end hydroxylated products. This, however, still may not account entirely for the lack of growth observed. It has been argued that degradation of xenobiotic compounds in the environment is hindered by the slow rate of evolution of key enzymes and pathways

111

for overcoming substrate range limitations (Ramos et al., 1987; Kumamaru et al., 1998), regulatory bottlenecks (Ramos et al., 1986; Pavel et al., 1994), mis-routing of novel substrates (Rojo et al., 1987) and toxicity of by-products (e.g. protoanemonin, Blasco et al., 1995). Such problems cannot explain the failure of the recombinant strains presented here to use 2-chlorotoluene as sole carbon source. Since the hybrid pathway is functional for degradation of 2-chlorotoluene when cell metabolism is maintained at the expense of a second carbon source (i.e. citrate), we conclude that degradation of this Cl-aromatic substrate is enzymatically possible, but physiologically unfavourable. In other words, the energy cost of building the functional and complete hybrid pathway may not be balanced by the metabolic return of 2-chlorotoluene metabolism. Consequently, it appears that such a bottleneck may not be solved by engineering an

Fig. 4. Monitoring metabolism of 2-chlorotoluene. The ability of genetically modified P. aeruginosa AH001 and P. aeruginosa AH002 strains to metabolize 2-chlorotoluene was tested as follows. Colonies of each of the strains under scrutiny, picked from a fresh 2-chlorobenzoate plate plus antibiotics, were resuspended in a small volume of sterile NaCl 1%, washed twice to eliminate all exudates which could be adhered to the cells and innoculated in 15 ml of mineral medium [K2HPO4 (7.0 g l − 1), KH2PO4 (3.0 g l − 1), MgSO.47H2O (1.0 g l − 1) and (NH4)2SO4 (1.0 g l − 1)] amended with 0.2% citrate as carbon and energy source. Alternatively, cultures were re-inoculated from cells pregrown in liquid 2-chlorobenzoate 5 mM medium, washed thoroughly and resuspended in 1% NaCl as before. In either case, the new cultures were adjusted to an initial absorbance at 600 nm of 0.05. Following 2 days of growth at 30°C with shaking, the cultures were exposed to saturating vapours of 2-chlorotoluene and further incubated for another 4 days. Control cultures involved non-recombinant, parental strains P. aeruginosa PA142 and P. aeruginosa JB2, as well as samples not exposed to 2-chlorotoluene vapours. The cultures were then centrifuged and the supernantants processed for gas chromatographymass spectrometry analysis (GC-MS). To this end, the 15 ml-supernantants were adjusted to pH 1.0 with concentrated HCl and extracted four times with diethyl ether. Water was removed from the organic phase with anhydrous Na2SO4, the hydrated salt removed by filtration and the ether evaporated under an N2 stream. The dry residue was resuspended in 75 ml of pyridine containing an internal standard of ethylvanilline (0.33 mg/ml). Aliquots (25 ml) of these samples were derivatised for 30 min at 80°C with 40 ml of bis(trimethylsilyl)-trifluoroacetamide (BSTFA) and injected in a Perkin– Elmer Autosystem apparatus, with helium as carrier gas. The column (SPB-1, Supelco) was 30 cm long, Ø 0.25 mm and a film thickness of 0.2 mm. Samples were run through a temperature gradient from 50–70°C (5 min) up to 280°C at a rate of 10°C per minute. The products in outcoming peaks were broken and ionized by electronic impact in a Perkin–Elmer Q-MASS apparatus and submitted to MS analysis. Assignation of MS signals to specific compounds involved the automated comparison of the distribution of masses with those of the Perkin– Elmer Q-MASS database and others produced on purpose with commercial standards (i.e. 2-chlorobenzoate). (A) GC-MS of standard 2-chlorobenzoate and detection of the compound in the culture medium of P. aeruginosa AH001. The upper panel of the figure shows the MS of a standard pure 2-chlorobenzoate sample. The medium panel shows the GC-MS analysis from one BSTFA-derivatized supernatant of a P. aeruginosa AH001 clone collected from a culture with mixed feeding of citrate and 2-chlorotoluene. The MS (m/z) analysis of the very predominant peak (lower panel) reveals its identity with 2-chlorobenzoate. (B) Analysis of intermediates in the degradation of 2-chlorotoluene by P. aeruginosa AH001 in citrate-fed cultures. The upper panel shows the whole GC profile of products found in the supernantant of another P. aeruginosa AH001 clone grown in citrate medium and exposed to saturating vapours of 2-chlorotoluene. Some peaks could be identified with confidence on the basis of their retention times on GC and the patterns of their MS. Those labelled 1–3 presented the shown MS (m/z) patterns and their predicted structures are indicated in each case. Other predominant peaks from products present in the culture medium gave TICs typical of hidroxylated aromatic rings, but the structure could not be determined accurately. Note that the ordinate in the TICs graphics corresponds to abundance (in percentage), while the abscises represent scan number in the upper row and retention time in the second row. For MS diagrams, the ordinates show the abundance (%) of each ionic species, while the abscises correspond to m/z values.

112

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113

enzyme or set of enzymes with higher activity on the substrate, but only by providing the biocatalyst with a separate energy source to compensate for the lack of metabolic return of the biodegradation process. This concept is at the basis of emerging phytoremediation and rhizoremediation strategies, in which the bacterial catalyst lives predominantly on plant root exudates while expressing one (Yee et al., 1998) or more (Brazil et al., 1995) genes encoding biodegradative activities. As a final, general consideration, even if the energy balance of the degradative process for a given compound were favourable, it is essential that expression of the introduced genes becomes properly coupled to the general physiology of the cells under operation conditions. Induction of promoter activity by specific effectors is frequenty overruled by physiological signals that inhibit transcription of individual genes if the metabolic status of the bacteria in not suitable (Cases and de Lorenzo, 1998). In this context, strong inducible promoters may hinder rather than favour the expression of genes and operons for metabolisation of some carbon sources. This realization opens new challenges for the outline of novel heterologous expression systems that are not detrimental for the physiology of the bacterial host. This also includes the utilisation of chemostats and non-homogeneous growth media for selecting strain variants that occupy distinct energetic niches (Rainey and Travisano, 1998). Some steps in that direction have been recently made in recombinant E. coli destined to perform in a bioreactor (Farmer and Liao, 2000) but should soon be extended to strains destined for environmental release and bioremediation (Cases and de Lorenzo, 1998).

Acknowledgements The authors are indebted to R.P. Hausinger and D.D. Focht for kindly providing strains, to Alicia Prieto and Angel Martı´nez’s group for help with GC-MS analyses and to D. Pieper for essential advice on the work and useful discussions. This research was supported by Contracts

BIO4-CT97-2040 and QLRT-99-00041 of the EU and by Grant BIO98-0808 of the Comisio´n Interministerial de Ciencia y Tecnologı´a. MAH was a predoctoral Fellow of the Spanish Ministry of Education and Culture.

References Abril, M.A., Michan, C., Timmis, K.N., Ramos, J.L., 1989. Regulator and enzyme specificities of the TOL plasmid-encoded upper pathway for degradation of aromatic hydrocarbons and expansion of the substrate range of the pathway. J. Bacteriol. 171, 6782– 6790. Blasco, R., Wittich, R.M., Mallavarapu, M., Timmis, K.N., Pieper, D.H., 1995. From xenobiotic to antibiotic, formation of protoanemonin from 4-chlorocatechol by enzymes of the 3-oxoadipate pathway. J. Biol. Chem. 270, 29229– 29235. Brazil, G.M., Kenefick, L., Callanan, M., Haro, A., de Lorenzo, V., Dowling, D.N., O’Gara, F., 1995. Construction of a rhizosphere Pseudomonads with potential to degrade polychlorinated biphenyls an detection of bph gene expression in the rhizosphere. Appl. Environ. Microbiol. 61, 1946– 1952. Brinkmann, U., Reineke, W., 1992. Degradation of chlorotoluenes by in vivo constructed hybrid strains: problems of enzyme specificity, induction and prevention of meta-pathway. FEMS Microbiol. Lett. 75, 81 – 87. Cases, I., de Lorenzo, V., 1998. Expression systems and physiological control of promoter activity in bacteria. Curr. Opin. Microbiol. 1, 303– 306. Cases, I., de Lorenzo, V., Pe´rez-Martı´n, J., 1996. Involvement of s54 in exponential silencing of the Pseudomonas putida TOL plasmid Pu promoter. Mol. Microbiol. 19, 7 – 17. de Lorenzo, V., Timmis, K.N., 1994. Analysis and construction of stable phenotypes in gram-negative bacteria with Tn5 - and Tn10 -derived minitransposons. Methods Enzymol. 235, 386– 405. de Lorenzo, V., Ferna´ndez, S., Herrero, M., Jakubzik, U., Timmis, K.N., 1993. Engineering of alkyl- and haloaromatic-responsive gene expression with mini-transposons containing regulated promoters of biodegradative pathways of Pseudomonas. Gene 130, 41 – 46. Farmer, W.R., Liao, J.C., 2000. Improving lycopene production in Escherichia coli by engineering metabolic control. Nature Biotechnol. 18, 533– 537. Ferna´ndez, S., de Lorenzo, V., Pe´rez-Martı´n, J., 1995. Activation of the transcriptional regulator XylR of Pseudomonas putida by release of repression between functional domains. Mol. Microbiol. 16, 205– 213. Haigler, B.E., Spain, J.C., 1989. Degradation of p-chlorotoluene by a mutant of Pseudomonas sp. strain JS6. Appl. Environ. Microbiol. 55, 372– 379.

M.-A. Haro, V. de Lorenzo / Journal of Biotechnology 85 (2001) 103–113 Harayama, S., Rekik, M., Wubbolts, M., Rose, K., Leppik, R.A., Timmis, K.N., 1989. Characterization of five genes in the upper-pathway operon of TOL plasmid pWW0 from Pseudomonas putida and identification of the gene products. J. Bacteriol. 171, 5048–5055. Hickey, W.J., Focht, D.D., 1990. Degradation of mono-, di-, and trihalogenated benzoic acids by Pseudomonas aeruginosa JB2. Appl. Environ. Microbiol. 56, 3842–3850. Knackmuss, H.J., 1996. Basic knowledge and perspectives of bioelimination of xenobiotic compounds. J. Biotech. 51, 287– 295. Kumamaru, T., Suenaga, H., Mitsouka, M., Watanabe, T., Furukawa, K., 1998. Enhanced degradation of polychlorinated biphenyls by directed evolution of biphenyl dioxygenase. Nature Biotech. 16, 663–666. Lau, P.C.K., Wang, Y., Patel, C ¸ .A., Labbe´, D., Bergeron, H., Brousseau, R., Konishi, Y., Rawlings, M., 1997. A bacterial basic region leucine zipper histidine kinase regulating toluene degradation. Proc. Natl. Acad. Sci. USA 94, 1453– 1458. Lehning, A., Fock, U., Wittich, R.-F., Timmis, K.N., Pieper, D.H., 1997. Metabolism of chlorotoluenes by Burkholderia sp. strain PS12 and toluene dioxygenase of Pseudomonas putida F1: Evidence for monooxygneation by toluene and chlorobenzene dioxygenases. Appl. Environ. Microbiol. 63, 1974– 1979. Manoil, C., Beckwith, J., 1985. Tnpho A: a transposon probe for protein export signals. Proc. Natl. Acad. Sci. USA 82, 8129– 8133. Panke, S., Sa´nchez-Romero, J.M., de Lorenzo, V., 1998. Engineering of quasi-natural Pseudomonas putida strains for toluene metabolism through an ortho-cleavage degradation pathway. Appl. Environ. Microbiol. 64, 748–751. Pavel, H., Forsman, M., Shingler, V., 1994. An aromatic effector specificiy mutant of the transcriptional regulator DmpR overcomes the growth constraints of Pseudomonas sp. strain CF.600 on para-substituted methylphenols. J. Bacteriol. 174, 7550–7557. Ramos, J.L., Stolz, W., Reineke, W., Timmis, K.N., 1986. Altered effector specificities in regulators of gene expression: TOL plasmid xylS mutants and their use to engineer expansion of the range of aromatics degraded by bacteria. Proc. Natl. Acad. Sci. USA 83, 8467–8471. Ramos, J.L., Wasserfakken, A., Rose, K., Timmis, K.N., 1987. Redesigning metabolic routes: manipulation of TOL plasmid pathway for catabolism of alkylbenzoates. Science 235, 593– 596. Rainey, P., Travisano, M., 1998. Adaptive radiation in a heterogenous environment. Nature (Lond.) 394, 69–72.

.

113

Reineke, W., 1998. Development of hybrid strains for the mineralization of chloroaromatics by patchwork assembly. Annu. Rev. Microbiol. 52, 287– 331. Robertson, J.B., Spain, J.C., Haddock, J.D., Gibson, D.T., 1992. Oxidation of nitrotoluenes by toluene dioxygenase: evidence for a monooxygenase reaction. Appl. Environ. Microbiol. 58, 2643– 2648. Rojo, F., Pieper, D.H., Engesser, K.H., Knackmuss, H.-J., Timmis, K.N., 1987. Assemblage of ortho cleavage route for simultaneous degradation of chloro- and methylaromatics. Science 238, 1395– 1398. Romanov, V., Hausinger, R.P., 1994. Pseudomonas aeruginosa 142 uses a three component ortho-halobenzoate 1,2-dioxigenase for metabolism of 2,4-dichloro- and 2-chlorobenzoate. J. Bacteriol. 176, 3368– 3374. Sa´nchez-Romero, J.M., Dı´az-Orejas, R., de Lorenzo, V., 1998. Resistance to Tellurite as a selection marker for genetic manipulations of Pseudomonas strains. Appl. Environ. Microbiol. 64, 4040– 4046. Shaw, J.P., Harayama, S., 1990. Purification and characterisation of TOL plasmid encoded benzyl alcohol dehydrogenase and benzaldehyde dehydrogenase of Pseudomonas putida. Eur. J. Biochem. 191, 705– 714. Timmis, K.N., Steffan, R.J., Untermann, R., 1994. Designing microorganisms for the treatment of toxic wastes. Annu. Rev. Microbiol. 48, 525– 557. van der Meer, J.R., de Vos, W.M., Harayama, S., Zehnder, A.J.B., 1992. Molecular mechanisms of genetic adaptation to xenobiotic compounds. Microbiol. Rev. 56, 677– 694. Vandenbergh, P.A., Olsen, R.H., Colaruotolo, J.F., 1981. Isolation and genetic characterization of bacteria that degrade chloroaromatic compounds. Appl. Env. Microbiol. 42, 737– 739. Williams, P.A., Shaw, L.M., Pitt, C.W., Vrecl, M., 1997. XylUW, two genes at the start of the upper pathway operon of TOL plasmid pWW0, appear to play no essential part in determining its catabolic phenotype. Microbiology 143, 101– 107. Yee, D.C., Maynard, J.A., Wood, T.K., 1998. Rhizoremediation of trichloroethylene by a recombinant, root-colonizing Pseudomonsa fluorescens strain expressing toluene orthomonooxygenase constitutively. App. Environ. Microbiol. 64, 112– 118. Zylstra, G.J., Gibson, D.T., 1989. Toluene degradation by Pseudomonas putida F1. J. Biol. Chem. 264, 14940– 14946. Zylstra, G.J., McCombie, R., Gibson, D.T.R., Finnette, B.A., 1988. Toluene degradation by Pseudomonas putida F1: genetic organization of the tod operon. Appl. Environ. Microbiol. 54, 1489– 1503.