Jan 13, 1995 - (in bijou bottles) and 10% (v/v) propene. The incubator was placed ..... otechnol 27:337-340. Whittenbury R, Phillips KC, Wilkinson JF (1970).
Appl Microbiol Biotechnol (1996) 44:582-588
© Springer-Verlag 1996
N . - Y . Z h o u • C. K. C h a n K w o C h i o n • D . J. L e a k
Cloningand expressionof the genes encodingthe propene monooxygenasefrom Xanthobacter, Py2
Received: 13 January 1995/Received revision: 7 June 1995/Accepted: 19 June 1995
Abstract Xanthobacter Py2 grows on propene as sole carbon source, converting propene to propene oxide (epoxypropane) using an alkene-specific monooxygenase, as the first step in catabolism. Four mutants, NZI-4, with a propene- propene oxide-- phenotype were isolated by 1-methyl-3-nitro-l-nitrosoguanidine mutagenesis or by enrichment with the suicide substrate vinylidene chloride, and were shown to have lost the ability to convert alkenes to epoxides. All four mutants were complemented by a number of clones of Xanthobaaer Py2 chromosomal DNA in the broad-host-range cosmid pLAFR5, some of which appeared to be non-overlapping. Representatives of the different clones obtained were transferred into Xanthobacter autotrophicus JW33 and one, pNY2, the most frequently isolated clone, was shown to express an inducible, fully functional propene monooxygenase. Subcloning revealed that all four mutants were complemented by a 2.4-kb EcoRI-PstIfragment situated at one end of the cosmid insert. However, activity in X. autotrophicus JW33 could only be expressed from pNY2, containing the complete insert (25 kb), suggesting a large operon or some form of long-range control. pNY2 failed to express in E. coli. In X. autotrophicus JW33 [-pNY2] at least three new polypeptides were evident after induction with propene compared with a control carrying only the cosmid pLAFR5.
Introduction Bacteria isolated for their ability to utilise low-molecular-mass alkenes (ethene, propene and 1-butene) as sole carbon source fall predominantly into three genera, N.-Y. Zhou • C. K. Chan Kwo Chion • D. J. Leak (L~) Centre for Biotechnology, Department of Biochemistry, Imperial College, London SW7 2AZ, UK. Fax 071 594-5207. E-mail d.leak @bc.ic.ac.uk
Mycobacterium spp., Nocardia spp. and Xanthobacter spp. (Ginkel et al. 1987). Mycobacterium spp. predominate when ethene is used as carbon source and are frequently isolated on propene, while in our experience Nocardia spp. are the most common isolates on 1butene. However, isolations using propene as sole carbon source yield the widest range of bacterial types including isolates of the gram-negative Xanthobacter spp. (Ginkel et al. 1987) There has been considerable interest in this group of organisms in recent years, as biocatalysts for the production of chiral epoxides for use as intermediates in asymmetric synthesis (Leak et al. 1992; de Bont 1993). This may be achieved either directly by epoxidation of terminal alkenes (i.e. with terminal double bonds) or by chiral resolution, exploiting the inherent enantioselectivity of epoxide degradation in some of these organisms (Weijers et al. 1988). These bacteria may also be useful for the biotreatment of toxic chlorinated alkenes (Hartmans et al. 1992). Unlike the alkane (e.g. methane, propane, octane and higher n-alkanes)-oxidising bacteria, which have also been shown to catalyse alkene epoxidation, most of the organisms isolated on low-molecular-mass alkenes possess an alkene-specific monooxygenase that converts 1,2-alkenes to their respective epoxides but does not catalyse C-H bond oxidation. Although some bacteria isolated on low-molecular-mass alkenes also grow on alkanes, this has been shown to be due to the presence of a second (alkane) monooxygenase, which does oxidise C-H bonds (de Bont et al. 1979). Bacteria expressing only the alkene monooxygenase have been shown to catalyse epoxidation of a wide range of substrates, and the former Nippon Mining Company (now Japan Energy Corporation) market chiral epoxides produced biocatalytically using a propene-utilising strain (B276) of Nocardia corallina. (Furuhashi 1992). However, regardless of the isolate used, the rates of epoxidation are low, even with the natural substrate. Ginkel et al. (1987) compared the
583
epoxidation rate of a range of isolates grown on their substrate of isolation, the highest being 81 nmol min-1 mg protein - ~ (based on total cell protein) recorded for Xanthobaaer Py2, the isolate used in this study. Furuhashi (1992) has recorded the rates of epoxidation of a range of alkenes with Nocardia corallina B276. Assuming that 1 mg dry weight of cells contains 0.5 mg total cell protein, the equivalent activity in B276 is only about 7 nmol min- 1 mg protein ~ (although there is no indication whether these are optimised rates.) The growth rates of the parental strains are also generally slow, typical doubling times on alkene substrates being 6-9 h. It would therefore be useful to be able to express the alkene monooxygenase to higher levels in an alternative host capable of more rapid growth rates. At a more fundamental level the cloning and sequencing of alkene monooxygenases will open up the opportunity to explore the molecular basis of substrate specificity and reaction specificity of this class of enzyme. This paper describes the cloning of the genes encoding alkene monooxygenase from a propene-utilising Xanthobacter sp. Py2 and expression of a fully functional monooxygenase in the reference strain Xanthobacter autotrophicus JW33. Although protein purification studies have concentrated mainly on the grampositive strains (i.e. Nocardia and Mycobacteria) the protein had not been purified in active form from any strain at the outset of these studies. Given the multicomponent nature of alkane monooxygenases and the preliminary evidence from studies of M. aurum E3 (Weber et al. 1992), it was reasonable to expect that alkene monooxygenase activity would require the coordinated expression of a number of genes. Without any indication of the number of genes involved or extensive sequence information to generate nucleotide probes, a mutation-complementation approach using broad-host-range cosmid vectors was considered to be the most direct route. Such an approach would not be feasible with Nocardia spp. or Mycobacterium spp. because of the lack of an established genetic system for the former and low transformation frequencies for the latter.
Materials and methods Strains and plasmids The following bacterial strains were used in this study: Xanthobacter Py2 propene + Tc s (Ginkel and de Bont 1986); Xanthobacter autotrophicus JW33 NCIMB 12468 (Wiegel et al. 1978); E. coli DH5c~ supE44AlacY169 (~b80 lacZ zXM15) hsdR17 recA1 endA1 gyrA96 thi-1 reIA1 (Hanahan 1985); E. coli HB101 supE44 hsdS20 (r~mB) fecAl3 ara-14 proA2 lacY1 galK2 rpsL20 xyl-5mtl-1 (Boyer and Roulland-Dussiox 1969). The following plasmids were used or created during this study: pLAFR5 a broad-host-range cosmid (RK2 replicon) Tc r, tra- mob + (Keen et al. 1988); pRK415 a broad-host range-plasmid (RK2 replicon) Tc r, tra- mob + (Keen et al. 1988); pRK2013 a helper plasmid (ColEI replicon) Km r, tra+ (Figurski and Helinski 1979); pNYI-4,
pLAFR5 containing inserts of Py2 chromosomal DNA (this study); pNY2A-G, subclones of pNY2 (this study); pNY2E 1-3, subclones of pNY2E (this study).
Growth media Media for growth of E. coli strains (LB medium, nutrient broth, SOB medium) were prepared according to the manufacturers' instructions or according to Sambrook et al. (1989). Xanthobacter Py2 was grown on an ammonia/mineral salts medium (AMS; Whittenbury et al. 1970) supplemented with 10% (v/v) propene, 0.01% (v/v) propene oxide or containing 0.5% (w/v) fructose. X. autotrophJcus JW33 was grown on AMS/fructose (0.5% w/v) or on AMS containing 0.1% (w/v) isopropanol. Flasks containing gaseous or volatile substrates were sealed with a Suba-seal serum cap (W.H. Freeman, Barnsley, UK) and plates for incubation in the presence of these substrates were placed in a standard incubator in a "pizza storer", provided with a gas inlet tube, or in an ENKAB-1 incubator (model 60, Elpas Ltd). Kanamycin (20 lag ml-1) or tetracycline (20 lag ml-1) for plasmid selection were added to the media as required.
Mutagenesis procedures Cells of Py2 grown overnight in AMS/propene as a 10% (v/v) subculture from a 48-h preculture in the same medium were subjected to mutagenesis with 170 lagml-1 1-methyl-3-nitro-l-nitrosoguanidine (MNNG) for 30 min at 30°C in 0.1 M citrate buffer pH 5.6 (Miller 1972). The survivors were grown on AMS/fructose plates. For selection of vinylidene-chloride-resistant mutants, cells grown for 48 h in AMS/fructose were spread on AMS plates and incubated (30 ° C) in an ENKAB-1 incubator (model 60, Elpas Ltd.) in the presence of 700 lal vinylidene chloride, 700 lal propene oxide (in bijou bottles) and 10% (v/v) propene. The incubator was placed in a fume hood for 2 weeks. Colonies arising after M N N G mutagenesis or vinylidene chloride selection were transferred to AMS/fructose plates and tested for growth on propene by replica plating on to AMS/propene plates. Propene- colonies were then tested for reversion and for whole-cell propene monooxygenase and propene oxide breakdown activity.
Preparation of a cosmid library The procedure for preparation of high-molecular-mass chromosomal DNA from Py2 was a modification of that described by Fulton et al. (1984). Cells growing in AMS propene were treated with 10 ~tg ml 1 ampicillin for 5 h prior to harvesting. Harvested cells were resuspended in 20 mM TRIS/HC1, 10 mM EDTA, 50 mM glucose, pH 8.0, treated with 1 mg ml 1 lysozyme at 37 ° C for 30 min and then frozen in solid CO2/ethanol. Sodium dodecyl sulphate(SDS) was added to the frozen cells to a final concentration of 1 % (w/v) and the mixture transferred to a 60 ° C water bath for 10 rain. The resulting lysate was treated with 50 lag m l - 1 DNase-free RNase A for 1 h at 37°C, then 50 lagm1-1 proteinase K for 1 h at 60°C. Following phenol/chloroform extraction, high-molecularmass DNA was precipitated with ethanol (Sambrook et al. 1989). After optimisation of conditions on a small scale, 200 lag of high-molecular-mass chromosomal DNA was partially digested with 0.8 unit Sau3AI (Northumbria Biologicals) at 37 ° C for 60 min, to a modal size of 30 kb, and dephosphorylated with alkaline phosphatase. Vector (pLAFR5) DNA was prepared from a 500-ml culture of E. coIi DH5e in LB medium using the Qiagen Maxi-prep protocol with a Qiagen-tip 500 (Qiagen Inc., Chatsworth, USA), digested sequentially with BamHI and ScaI, extracted with phenol/chloroform and precipitated with ice-cold ethanol. After
584 resuspending in 10 mM TRIS/HC1 pH 8.0, vector and chromosomal fragments were ligated at 16° C for 18 h with a ninefold molar excess of cosmid vector and packaged into )~ heads using Gigapack Gold (Stratagene) packaging extracts. The resulting library was titred using E. coli DH5c~ as host and subsequently amplified in the same host.
Triparental mating The E. coli donor containing recombinant plasmid/cosmid and helper strain HB101 [pRK2013] was grown separately in LB medium for 5 h while mutants of Py2 or X. autotrophicus JW33 were subcultured overnight in AMS/fructose from a 48-h culture in the same medium. Donor, helper and recipient cells were centrifuged, washed and resuspended in AMS and then mixed in appropriate ratios on a 0.2-gin filter disc (Millipore) placed on an AMS/fructose plate. After overnight conjugation at 30°C the cells were resuspended in AMS and spread on selective media.
Assays of propene oxide formation and degradation Assays were done on resting cell suspensions of Xanthobacter Py2, the mutants NZl 4 and X. autotrophicus JW33 transconjugants. Cells were initially grown on AMS/propene, AMS/propene/propene oxide and AMS/isopropanol/propene respectively and harvested in late exponential phase. Propene oxide formation (in 25 mM glycine/ NaOH buffer, pH 9.0) and degradation (in 25 mM phosphate buffer, pH 7.0) were assayed in 7-ml sealed (Suba-seal) flasks containing 1 ml cell suspension with an absorbance at 540 nm of 10. Propene oxide was measured in 5-pl liquid samples removed through the Sub-seal, by flame ionization detection gas chromatography on Porapak Q, 80-100 mesh, at 180° C with nitrogen carrier gas at 40 ml rain -1. Propene oxide production activities are expressed in nmol min 1 mg protein-1, where the amount of protein refers to total whole cell protein measured by the method of Lowry et al. (1951) on cell suspensions pretreated by boiling for 5 min in 0.5 M NaOH (final concentration).
Results
Isolation and characterisation of mutants Conditions for M N N G mutagenesis were optimised to achieve approximately 50% survival. A total of 1000 survivors were screened for growth on propene, yielding a single propene- mutant, NZl, with a reversion frequency of less than 10-9. To improve the frequency of mutant isolation a variation on the enrichment technique used by Hartmans et al. (1988) was followed. This used vinylidene chloride (boiling point 30-32 ° C) in place of vinyl chloride (gaseous at room temperature) because of the greater ease of handling. Previous experiments had established that vinylidene chloride was an effective suicide substrate under the conditions used, selectively killing cells expressing propene monooxygenase, with reduced toxicity to cells grown on AMS/fructose. Selection in the presence of propene, propene oxide and vinylidene chloride, as described, yielded in excess of 500 mutants with a propenepropene oxide ÷ phenotype. The frequency of isolation
(2 x 10- 6) was higher than expected for a spontaneous mutation, suggesting that vinylidene chloride, propene oxide or both were also serving as mutagens. Ten mutants, NZ2-11 were picked and shown to have reversion frequencies of 10 - a to less than 10 - 9 . In whole-cell assays, mutants N Z l M , which had been grown on a mixture of propene oxide and propene (for induction), had no detectable propene monooxygenase activity but degraded propene oxide at the same rate as the wild type (approximately 50 nmol min 1 mg protein-i). When grown under the same conditions the wild-type strain Py2 had a propene monooxygenase activity of 20 nmol min - 1 mg protein - 1. Preparation of a cosmid library The broad-host-range cosmid pLAFR5 employs a double cos site with an internal blunt-end restriction site (ScaI). Digestion with B a m H I + ScaI yields cosmid arms that can be used without separate purification; concatemerisation and the consequent risk of packaging of multiple copies of the vector lacking inserts is avoided by the inclusion of a high concentration of ATP during the subsequent ligation step. This suppresses blunt-end ligation at the ScaI site. (Feretti and Sgaramella 1981). pLAFR5 is large for a cosmid vector but should be able to accommodate inserts of approximately 30 kb. Using the procedure described, sufficient high-molecular-mass chromosomal DNA was obtained from 6 x 50-ml cultures of Py2. Partial digests were not size-fractionated because of the losses encountered, but were treated with phosphatase to avoid insert-insert ligation. Successful ligation of cosmid and insert after 18 h was confirmed by gel electrophoresis (0.4% w/v agarose) and the product packaged into 7 heads as described. Estimating the titre of the library with E. coli DH5c~ demonstrated that a total of 10 000 Tc r colonyforming units had been obtained from approximately 0.2 mg packaged DNA. This was amplified in DH5c~ to a concentration of 5.4 x 10 l° cells/ml and preserved at - 70°C in 20% (v/v) glycerol. The cloning site in pLAFR5 is contained within the lacZ' gene facilitating screening for inserts by blue/white selection on isopropyl thiogalactoside/5bromo-4-chloro-3-indolyl /~-D-galactoside LB plates. This revealed that, despite the precautions taken, approximately 10% of the clones lacked inserts. Ten white (insert-containing) colonies were picked at random and their insert DNA analysed by Hind I I I + EcoRI digestion of small-scale plasmid preparations (Sambrook et al. 1989). All ten inserts were larger than 21 kb with an average insert size of 23.5 kb. According to the relationship of Clarke and Carbon (1976) and assuming the Xanthobacter Py2 chromosome is 4000 kb in size, this implied that a complete genomic library for screening purposes should be
585
contained in 900 library clones (this takes into account the 10% without inserts).
Complementation of propene- mutants The cosmid library was conjugated en masse in four triparental matings to the mutants N Z l - 4 and transconjugants selected on AMS/fructose/tetracycline plates and AMS/propene/tetracycline plates. Transconjugants were obtained at a frequency of 10- L 1 0 - ~ per recipient and complementation (observed after 1 week incubation) obtained at a frequency of 8 × 10-4-1 x 10- 3/transconjugant, which is only slightly lower than the predicted value. Ten of the complemented colonies (three each from NZ1 and NZ2, two each from NZ3 and NZ4) were picked and grown in AMS/propene/tetracycline and the cosmids were isolated by standard mini-preparative methods. Restriction patterns after EcoRI digestion revealed that seven of the inserts were identical while three, isolated from transconjugants derived from three different mutants (NZ1, 2 and 3) were apparently unique. One representative cosmid of each type was retained for further analysis and renamed pNYI-4 with pNY2 as the sole representative of the most frequent isolate. These were transformed back into E. coIi DH5c~. Sixteen sets of triparental matings were then done to test all combinations of cosmid and mutant. This demonstrated that all of the cosmids pNYI-4 complemented all four mutants NZI-4 (100% complementation per transformant).
Cosmid analysis The inserts in cosmids pNYI-4 were mapped by single and all combinations of double digests with HindIII, BamHI, EcoRI and ScaI (Fig. 1). Surprisingly, this did not reveal any common regions that would have provided a focus for the subcloning strategy. Hybridization studies by Southern blotting (Sambrook et al. 1989), Fig. 1 Restriction maps and relative sizes of the inserts from cosmids pNY1-4. Restriction sites within the inserts are indicated by the solid lines. Restriction sites within the vector or at the junction between vector and insert (multiple cloning sites) are indicated by the dashed lines
using the three fragments generated by BamHI digestion of the pNY2 insert as probes, also failed to reveal any homology. Therefore, on the basis of frequency of occurrence it seemed prudent to proceed with pNY2. This rationale was supported by the subsequent observation that pNY2 expressed a fully functional propene monooxygenase in transconjugants of the reference strain X. autotrophicus JW33 (see later). Cosmid pNY2 was subcloned as outlined in Fig. 2a. Initially vector pRK415 was used for subcloning but some of the constructs proved to be unstable [instability has also been reported by Keen et al. (1988)]. Therefore all of the subcloning was done in pLAFR5 either by restriction and self-ligation or by isolating fragments (Geneclean Biol01) and cloning them into the multiple cloning site of pLAFR5 (e.g. pNY2E was created by HindIII digestion of pNY2 and subsequent religation, relying on the fact that internal ligation would be more favourable than fragment rcinsertion; pNY2D was created by isolating the smaller fragment released in the HindlII digest and ligating it into the multiple cloning site of pLAFR5. All subclones were checked for size before proceeding). All of the subclones were tested for complementation of the mutants NZ1-4 and expression in X. autotrophicus JW33. Of the subclones containing inserts A-G only pNY2E and C complemented the mutants. Further mapping (Fig. 2b) of pNY2E revealed a suitably placed PstI site to enable subcloning of a 2.3-kb region originating from one end of the original clone. This subclone also complemented all four mutants. The location of the complementing fragment at the right-hand end of fragment E (Fig. 2a) was surprising as the BamHI fragment G, which incorporates all of this 2.3-kb region except approximately 300bp at the right-hand end, failed to complement any of the mutants.
Heterologous expression of PMO Expression of PMO activity from the complementing constructs pNY1-4 in E. coli DH5c~ was examined in
Hind III \ BarnHI EcoRI
Hind III \
Barn HI
,
Hind III Sca I Hind III
Sca I
I
\1
IpNY1 I Barn HI Eco RI
I
I
Scal Hindlll EcoRI
3
!
Hindlll
",/
l
/
p N Y 2 Insert 25.7 kb
BamHl
Hindlll
Scal
\ /
Eco RI
pNY3 Eco RI
p N Y 4 Insert 22.0 kb
[
5kb
J
Multiple cloning sites
Eco RI
