Analysis of Pseudomonas putida alkane-degradation gene clusters and flanking insertion sequences: evolution and regulation of the alk genes

Jan B. van Beilen1, Sven Pankea,1, Sacha Lucchinib,1, Alessandro G. Franchini1, Martina Röthlisberger1 and Bernard Witholt1

Institute of Biotechnology, Swiss Federal Institute of Technology (ETH), ETH-Hönggerberg, CH-8093 Zürich, Switzerland1

Author for correspondence: Jan B. van Beilen. Tel: +41 1 6333444. Fax: +41 1 6331051. e-mail: vanbeilen{at}biotech.biol.ethz.ch


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
The Pseudomonas putida GPo1 (commonly known as Pseudomonas oleovorans GPo1) alkBFGHJKL and alkST gene clusters, which encode proteins involved in the conversion of n-alkanes to fatty acids, are located end to end on the OCT plasmid, separated by 9·7 kb of DNA. This DNA segment encodes, amongst others, a methyl-accepting transducer protein (AlkN) that may be involved in chemotaxis to alkanes. In P. putida P1, the alkBFGHJKL and alkST gene clusters are flanked by almost identical copies of the insertion sequence ISPpu4, constituting a class 1 transposon. Other insertion sequences flank and interrupt the alk genes in both strains. Apart from the coding regions of the GPo1 and P1 alk genes (80–92% sequence identity), only the alkB and alkS promoter regions are conserved. Competition experiments suggest that highly conserved inverted repeats in the alkB and alkS promoter regions bind AlkS.

Keywords: alkane degradation, insertion sequence, chemotaxis, regulation

The EMBL accession numbers for the sequences reported in this paper are AJ245436 [P. putida (oleovorans) GPo1 alk gene clusters and flanking DNA], AJ233397 (P. putida P1 alk gene clusters and flanking DNA), AJ249793 (P. putida P1 nahKJ genes), AJ249825 [P. putida (oleovorans) GPo1 16S RNA gene] and AJ271219 (P. putida P1 16S RNA gene).

a Present address: DSM Research, Postbus 18, 6160 MD Geleen, The Netherlands.

b Present address: Institute of Food Research, Norwich Research Park, Colney, Norwich NR4 7UA, UK.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Even though the ability to grow on alkanes is a very common phenomenon, most biochemical and genetic studies have focused on a limited number of prototype systems. An example is the Pseudomonas putida GPo1 (commonly known as Pseudomonas oleovorans GPo1=TF4-1L=ATCC 29347) alkane hydroxylase system (Baptist et al., 1963 ), which can be used to carry out a wide range of stereo- and regioselective oxidation reactions (Witholt et al., 1990 ). In the alkane degrader Pseudomonas putida P1, closely related genes were found. Recent PCR studies (Smits et al., 1999 ) and genome sequencing have shown that many strains contain homologues of the GPo1 alkane hydroxylase, usually encoded on the chromosome. In contrast, the GPo1 alkane degradation (alk) genes are grouped in two clusters located on a large plasmid, named OCT (Chakrabarty et al., 1973 ). Mapping experiments had indicated that the two loci (Benson et al., 1977 ; Fennewald & Shapiro, 1977 ) are transcribed towards each other and are separated by about 40 kb (Fennewald et al., 1979 ; Owen, 1986 ). The G+C content of the alk genes is much lower than that of the host strain and the OCT plasmid (Fennewald et al., 1978 ; Kok et al., 1989b ; Panke et al., 1999b ), suggesting that these genes are part of a 55 kb mobile element, which integrated in the OCT plasmid. This appears to be a common finding for catabolic genes, and many catabolic transposons have now been described (Tan, 1999 ).

Witholt and co-workers cloned the two alk clusters in pLAFR1 as 18 kb and 16·9 kb EcoRI fragments to create plasmid pGEc47 (Eggink et al., 1987a ). Subsequent minicell experiments and partial sequence analysis showed that the 18 kb fragment contains an operon named alkBFGHJKL, which encodes two components of the alkane hydroxylase system and enzymes involved in further metabolic steps (Eggink et al., 1987b ; Kok et al., 1989a , b ; van Beilen et al., 1992a ). The 16·9 kb fragment was found to encode the third component of the alkane hydroxylase system: rubredoxin reductase, and AlkS, which regulates expression of the alkBFGHJKL operon (Kok et al., 1989b ; Eggink et al., 1990 ; Panke et al., 1999b ; Canosa et al., 2000 ), as well as the alkST genes (Canosa et al., 1999 , 2000 ) (see Fig. 1). Only alkB, alkG and alkS (Kok et al., 1989a ; van Beilen, 1994 ; van Beilen et al., 1992a ) cannot be complemented by genes present on the chromosome of GPo1.



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 1. Metabolic pathway of alkane degradation, role and cellular localization of Alk proteins and regulation of the alk genes. The alkBFGHJKL operon encodes the alkane hydroxylase (AlkB), two rubredoxins (AlkF and AlkG), an alcohol and an aldehyde dehydrogenase (AlkJ and AlkH, respectively), an acyl-CoA synthetase (AlkK) and an outer-membrane protein of unknown function (AlkL) (Kok et al., 1989a , b ; van Beilen et al., 1992a ). The alkST locus encodes the third component of the alkane hydroxylase system, rubredoxin reductase (AlkT), and AlkS, which positively regulates expression of the alkBFGHJKL operon (Eggink et al., 1990 ; Panke et al., 1999b ) and the alkST genes (Canosa et al., 2000 ). AlkN is a putative chemotactic transducer for alkanes.

 
To obtain more evidence for the existence of an alk transposon, and to investigate the possibility that additional non-essential functions involved in alkane degradation might be encoded outside the regions that had been sequenced to date, we completed the sequence of the two EcoRI fragments in pGEc47 and cloned additional DNA that links the two EcoRI fragments. Comparison of the P. putida P1 alkBFGHJ genes (Smits et al., 1999) with the GPo1 alkBFGHJKL genes showed that the homology between the alkB promoter regions ends with an inverted repeat. To better understand the significance of this and other sequence features, we also cloned and sequenced the remaining P. putida P1 alk genes and flanking DNA. Analysis of sequences flanking the alk gene clusters in both strains revealed the presence of several complete and incomplete insertion sequences (IS), which are likely to have played a role in mobilizing the alk gene clusters.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Bacterial strains and growth conditions.
P. putida GPo1 [commonly known as P. oleovorans GPo1=TF4-1L=ATCC 29347 (Schwartz, 1973 )] is a prototroph strain that carries the OCT-plasmid. P. putida P1 is a prototroph strain isolated from a pentane enrichment culture (Smits et al., 1999 ). E. coli DH10B (Gibco-BRL) was used for cloning and the production of plasmid DNA for sequencing. E. coli W3110 [F- {lambda}- IN(rrnDrrnE)1] is a prototroph K-12 strain (Bachmann, 1987 ) used for chemotaxis studies. E. coli JM101(alkS* alkBpxylE) (Panke et al., 1999a ) was used for regulation studies. All strains were grown in E2 medium (Lageveen et al., 1988 ) supplemented with the appropriate carbon source, or Luria–Bertani broth (Luria et al., 1960 ). E. coli DH10B was transformed by electroporation (Dower et al., 1988 ). To select E. coli transformants, antibiotics were used at the following concentrations (µg ml-1): ampicillin, 200; tetracycline, 12·5; kanamycin, 50.

DNA methods.
Restriction enzymes, T4-DNA ligase and dideoxynucleotides were obtained from Roche Diagnostics. PCR reactions were carried out as described by Innis et al. (1990) , using a Perkin Elmer GeneAmp PCR System 9600, and oligonucleotides were synthesized by Microsynth. Taq DNA polymerase and the Expand Long Template PCR system were obtained from Promega and Roche Diagnostics, respectively. For cloning, PCR products were digested with appropriate restriction enzymes, purified over a 1% agarose gel in TBE buffer and isolated by electroelution (Sambrook et al., 1989 ). PCR fragments were cloned between the same or compatible sites of pGEM-7Zf(+) (Promega) or pZERO-2 (Invitrogen). Plasmid DNA was isolated with the Roche Diagnostics High Pure plasmid isolation kit or, for large plasmids, as described by Kado & Liu (1981) . PFGE (Bustamante et al., 1993 ) was carried out using the Bio-Rad Chef-DR II Electrophoresis Cell. Chromosomal DNA was isolated with the Qiagen Genomic DNA kit.

The Roche Diagnostics DIG kit was used for Southern blots, using CSPD [disodium 3-(4-meth-oxyspiro {1,2-dioxetane-3,2'-(5'-chloro) tricyclo (3.3.1.13,7) decan}-4-yl)phenyl phosphate] as the chemiluminescent substrate. To prepare probes, appropriate DNA fragments were purified using a 1% agarose gel, electroeluted and labelled (DIG labelling kit; Roche Diagnostics). Hybridizations were carried out using standard hybridization buffer without formamide at 64 °C.

Cloning and sequencing of alk genes and flanking DNA.
To construct enriched gene banks, suitable restriction fragments were identified by Southern blotting using previously cloned GPo1 alk genes or flanking DNA as probes. Restriction fragments around the target size were cut from a preparative agarose gel, isolated by electroelution, ligated between the appropriate sites of pGEM-7Zf(+) or pZERO-2 and transformed into E. coli DH10B. Transformants containing the desired DNA fragment were identified by colony blotting using Roche Diagnostics nylon membrane for colony and plaque hybridization and the Roche Diagnostics DIG kit.

To clone DNA downstream of alkL in GPo1, a 0·8 kb SalI fragment located between alkL and the end of the EcoRI fragment was DIG-labelled and used as probe to identify an overlapping 3·0 kb NheI–HindIII fragment. A DIG-labelled GPo1 alkST fragment was used as probe to clone the P1 alkST genes as an EcoRI–HindIII and a HindIII fragment. Sequencing showed that an additional 400 bp HindIII fragment was missing from alkS, which was obtained by PCR using primers P1-alkSTBam (GCTCGCGGTGGATCCTCAGGTC) and P1-alkSTSac (GCGGCGGAGCTCGCGGTGGATGCTCAGGTC), and sequenced. The previously cloned EcoRI–HindIII fragment containing the P1 alkBFGHJ' genes (Smits et al., 1999 ) forms a contiguous sequence with the EcoRI–HindIII fragment containing most of P1 alkS (the EcoRI sites upstream of alkS and alkB are the same). The Expand Long Template system was used with primers P1-J-Xba (GGTCTAGAGTGGCGCTGCGTGCCCGTAAGC) and P1-STBamSac (GCGGCGGAGCTCGCGGTGGATCCTCAGGTC) to obtain DNA segments downstream of alkJ and alkT. (This PCR was carried out based on the erroneous assumption that the P1 alk genes are organized as in GPo1, where alkST is located downstream of alkL. A 7·2 kb PCR fragment was obtained because almost identical copies of insertion element ISPpu4 flank the P1 alk genes, and cause a crossover reaction in the PCR.) In addition, a 3·1 kb EcoRI fragment containing ISPpu4.2, and a 7·5 kb EcoRI–SalI fragment containing the P1 alkL gene, ISPpu4.1, and DNA further downstream, were cloned.

A combination of techniques was used to sequence the cloned DNA fragments. Subclones for sequencing of the 8·0 kb DNA segment downstream of alkT were prepared by making nested deletions with Exonuclease III (Henikoff, 1984 ). The other DNA fragments were sequenced by subcloning selected restriction fragments, by making directed deletions and by primer walking. Cycle sequencing reactions were carried out using the Amersham Thermosequenase kit with deaza-GTP and IRD-800-labelled primers obtained from MWG-Biotech. Sequencing was carried out on a Li-Cor 4000L sequencer. DNA and protein sequences were analysed using Lasergene Navigator (DNASTAR). BLAST sequence comparisons were carried out at NCBI (Altschul et al., 1990 ), using standard settings or match value, mismatch value and gap extension penalty -r 5, -q -4 and -E 4, respectively, to detect low levels of sequence homology. 16S rRNA genes were amplified with primers 16F27 and 16R1525, and sequenced with 16F355, 16R519 and 16R1488 (Hauben et al., 1997 ).

Construction of plasmids and strains for alkB and alkS promoter studies.
The three inverted repeats (IR-B, IR-S and IR-syn) were synthesized as linkers with EcoRI and HindIII sticky ends (see Fig. 3c) and ligated to plasmid pGEM-7Zf(+) (Promega) digested with EcoRI and HindIII. Plasmid DNA isolated from E. coli DH10B transformants, selected by standard blue/white screening, was sequenced to check the inserts. E. coli JM101(alkS* alkBpxylE) was transformed with the three IR-vectors, pGEM-7Zf(+) and pBG201. M9 medium (100 ml), containing 0·2% glucose and 0·001% thiamine, was inoculated with 2 ml overnight LB-medium cultures. At an OD450 of 0·5 the cultures were induced with 0·02% dicyclopropylketone (Grund et al., 1975 ). Catechol-2,3-dioxygenase (XylE) activity was assayed (Nozaki, 1970 ) 2 h after induction.



View larger version (51K):
[in this window]
[in a new window]
 
Fig. 3. (a) Alignment of the GPo1 and P1 alkB promoter regions and the GPo1 alkN promoter region. The underlined sequences marked -35 and -10, and the arrow labelled +1, refer to the promoter and mRNA start site identified previously (Canosa et al., 2000 ; Kok et al., 1989b ). The translation start sites of AlkB and AlkN (predicted) and the stop codon of the short AlkB' peptide encoded by the DNA segment upstream of the alkN coding region are also underlined. Box A indicates that this and DNA further upstream are homologous to an internal fragment of IS1384. Box B indicates the extended homology between the P1 alkB and GPo1 alkN promoter regions, compared to the GPo1 alkB promoter region. The horizontal arrows (C) mark the 20 bp inverted repeat proposed as AlkS binding site. Box D marks the end of the homology between the sequence upstream of the alkN coding region and the alkB promoter regions, about 80 bases within the alkB gene and 25 bases upstream of the alkN start codon. RBS, Ribosome-binding site. (b) Alignment of the GPo1 and P1 alkS promoter regions. The underlined sequences marked -35 and -10, and the arrows labelled +1, refer to the {sigma}S- and AlkS-dependent promoters identified previously (Canosa et al., 1999 , 2000 ). Box A indicates that this DNA and sequences further upstream are homologous to the end of IS1384. Box B marks the start of the homology between the GPo1 and P1 alkS promoter regions. The horizontal arrows (C) mark the 18 bp perfect inverted repeat proposed as AlkS-binding site. The start codon of alkS is underlined. (c) Sequence of inverted repeat primers used to construct pGEM-7Zf(+) IR-B, IR-S and IR-syn. The horizontal arrows mark the inverted repeats.

 
Accession numbers and nomenclature.
Sequences determined in this study were deposited in the EMBL database and are available under accession numbers AJ245436 [P. putida (oleovorans) GPo1 alk gene clusters and flanking DNA], AJ233397 (P. putida P1 alk gene clusters and flanking DNA), AJ249793 (P. putida P1 nahKJ genes), AJ249825 [P. putida (oleovorans) GPo1 16S RNA gene] and AJ271219 (P. putida P1 16S RNA gene). The insertion sequences were named as proposed by the curators of the insertion sequence database (http://www-is.biotoul.fr/is/is_about.html, 18 October 2000). In the absence of a standardized nomenclature for transposons, the P. putida P1 alk transposon was named TnPpu-alk1.


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
P. putida (oleovorans) GPo1 utilizes medium-chain-length alkanes as its sole carbon and energy source and has been studied in detail with respect to the genetics and enzymology of alkane oxidation (reviewed by van Beilen et al., 1994 ) (Fig. 1). We cloned and sequenced DNA segments flanking the alk genes of GPo1 to identify additional genes involved in alkane degradation and to investigate the possibility that the alk genes of this strain are organized in a transposon-like structure. To be able to judge the significance of sequence features and gene organization of the GPo1 alk genes, we also cloned and sequenced the related alk genes of P. putida P1.

Taxonomical analysis of Pseudomonas strains GPo1 and P1
The strain commonly known as P. oleovorans GPo1 (=TF4-1L=ATCC 29347) was first described in 1963 (Baptist et al., 1963 ) and tentatively identified as a strain of P. oleovorans. After strain improvement for the production of epoxides (Schwartz & McCoy, 1973 ) it was submitted to ATCC as P. oleovorans TF4-1L, and is available as ATCC 29347. However, in the extensive phylogeny paper of Stanier et al. (1966) , the original isolate was listed as strain 266 and already classified as a P. putida biotype A (Chakrabarty et al., 1973 ). Strain 266 was also submitted to ATCC and is available but unlisted (ATCC 17633). As expected, ATCC 17633 is able to grow on n-octane and the 16S sequences of GPo1 and ATCC 17633 are identical. Sequencing of the complete GPo1 16S sequence showed that it is closely related to the 16S sequences of P. putida F1 (gb:L37365) and the P. putida biotype A type strain DSM 291T (5 and 12 differences, or 99·7 and 99·2% sequence identity, respectively). It is more distantly related to the 16S sequence of other Pseudomonas species (>30 differences, or <98% sequence identity), including the type strain of P. oleovorans, DSM 1045T, which was first described in 1941 (Lee & Chandler, 1941 ). This confirms the earlier classification of GPo1 by Stanier et al. (1966) . In this paper we refer to GPo1 as P. putida (oleovorans) GPo1.

P. putida P1 was originally isolated from a pentane enrichment culture inoculated with soil from a gas station in Groningen (Bioclear). It was classified as a P. putida biotype A strain by NCIMB. Sequencing of the complete P1 16S sequence showed that it is closely related to the GPo1 16S sequence (three differences, or 99·8% sequence identity).

Cloning and sequencing of DNA flanking the P. putida (oleovorans) GPo1 alk gene clusters
Witholt and co-workers cloned, and partially sequenced, an 18 kb EcoRI fragment containing the alkBFGHJKL genes (9·1 kb sequenced), and a 16·9 kb EcoRI fragment containing the alkST genes (4·7 kb sequenced) (Eggink et al., 1987a , 1990 ; Kok et al., 1989a , b ; Panke et al., 1999b ; van Beilen et al., 1992a ). We completed the sequence of both EcoRI fragments (Fig. 2). The region downstream of alkL in the 18 kb EcoRI fragment was found to encode a putative transposase of which a few N-terminal codons were missing due to the EcoRI cloning site. A 3·0 kb NheI–HindIII fragment which was cloned to complete the putative transposase (tnpA1) sequence was found to overlap with the sequence downstream of alkT on the 16·9 kb EcoRI fragment: a 651 bp EcoRI fragment sufficed to join the 16·9 and 18 kb EcoRI fragments. The resulting restriction map is in agreement with all restriction sites mapped in Southern blots (data not shown). This confirmed the hypothesis that the alk gene clusters are located back to back on the OCT plasmid. However, the two clusters are separated by only 9·7 kb (Fig. 2), instead of the 40 kb suggested by crossover experiments (Fennewald et al., 1979 ).



View larger version (18K):
[in this window]
[in a new window]
 
Fig. 2. Sequence analysis of the P. putida (oleovorans) GPo1 and the P. putida P1 alk genes and flanking DNA. The scale is in bp. Arrows represent potential coding regions. Black arrows represent genes involved, or presumably involved, in alkane degradation and vertically hatched arrows represent complete or partial transposase genes. Horizontally hatched arrows (also in combination with vertically hatched) indicate that the ORFs are incomplete or interrupted by stop codons or frameshifts. Black bars correspond to (incomplete) insertion sequences. ISPpu1–5 are new IS elements named in this study. The other black bars represent incomplete IS elements, with the name of the most closely related IS element shown between parentheses. The orientation of the IS elements is marked by L and R for left and right end. Bars linking the GPo1 and P1 DNA segments indicate homologous regions. DR, direct repeats formed by ISPpu4.1 and ISPpu4.2. See Table 1 and Fig. 1 for additional information on coding regions.

 

View this table:
[in this window]
[in a new window]
 
Table 1. Proteins encoded by the alk gene clusters of P. putida (oleovorans) GPo1 and P. putida P1, and flanking DNA

 
A chemotactic transducer for alkanes?
Sequence analysis of the 9·7 kb DNA segment separating the GPo1 alk gene clusters showed that a protein encoded 2 kb downstream of alkL shows up to 30% amino acid sequence identity with methyl-accepting chemotactic transducers, recognizing, for example, naphthalene (Grimm & Harwood, 1999 ) or amino acids (Taguchi et al., 1997 ). Like the alk genes, the chemoreceptor gene has a much lower G+C content than the flanking DNA and the OCT plasmid as a whole (Table 1). In addition, the 230 bp region directly upstream of the ATG start codon has 56% DNA sequence identity with the GPo1 alkB promoter region and the first 76 bases of the alkB gene, which suggests that the chemoreceptor gene (alkN) plays a role in alkane degradation (Fig. 3). Unfortunately, GPo1 is not very motile and could not be used for chemotaxis experiments, whilst P1 only contains a remnant of the alkN gene (see below).

Cloning and sequencing of the P. putida P1 alk gene clusters
Previously, a 7·5 kb EcoRI–HindIII fragment encoding the P1 alkBFGHJ genes was cloned, sequenced and shown to complement an E. coli recombinant containing all genes necessary for growth on n-octane except alkB (Smits et al., 1999 ) or alkG (rubredoxin 2) (J. B. van Beilen, unpublished data). The remaining alk genes and flanking DNA were cloned as overlapping DNA fragments using DIG-labelled probes (the GPo1 alkST or alkJKL genes) or PCR. Attempts to localize the P1 alk genes were not conclusive. P1 contains a 114±7 kb plasmid, as demonstrated by PFGE experiments (data not shown), but subsequent Southern blots proved that the alk genes are not located on this plasmid. We were unable to detect mobilization of the P1 alk gene clusters to other P. putida strains.

Sequence analysis and comparison of the P. putida (oleovorans) GPo1 and P. putida P1 DNA segments
The GPo1 and P1 sequences (34902 and 23312 bp, respectively) were analysed for coding regions, homology to each other and homology with database sequences. Table 1 lists all ORFs that have significant homology with protein database sequences and other ORFs over 100 aa. For each of the ORFs, length, position, G+C content of coding DNA, putative ribosome-binding site and best score in BLASTX searches are given. The data are graphically presented in Fig. 2. In some cases ORFs were extended or shortened to alternative start codons if BLASTX searches indicated that the coding region should be longer or shorter and/or if no ribosome-binding site could be detected.

Although the alkBFGHJKL and alkST gene clusters are organized similarly in both strains (Fig. 2), an interesting difference between GPo1 and P1 is the relative position of the gene clusters: in P1, alkST is located upstream of alkBFGHJKL. On average, the level of DNA sequence identity between the GPo1 and P1 alk genes is about 80% for the alkBFGHJKL genes, and 92% for the alkST genes, but only within the ORFs. The intergenic regions could not be aligned.

Most of the GPo1 and P1 Alk proteins show similar levels of sequence identity (Table 1). However, specific regions within the proteins, and the corresponding DNA sequences are not (well) conserved. The C-terminal 50 bp of the GPo1 and P1 alkB genes could not be aligned. In the case of the GPo1 alkB gene, this segment was shown previously to be non-essential to monooxygenase function; it could be replaced by the lacZ gene without losing AlkB activity (van Beilen et al., 1992b ).

The GPo1 rubredoxins AlkF and AlkG are unusual compared to rubredoxins isolated from anaerobic organisms. AlkF consists of an N-terminal rubredoxin-like domain and a C-terminal 80 bp extension that has no homology with database sequences. Similarly, AlkG consists of an N-terminal and a C-terminal rubredoxin-like domain, linked by a 70 bp sequence that has no homology with database sequences. Only the C-terminal rubredoxin-like domain of AlkG is essential for the alkane hydroxylase system (Kok et al., 1989a ). In P1, the C-terminal extension of AlkF and the AlkG linker are not well conserved. However, the N-terminal rubredoxin-like domains of AlkF and AlkG (AlkF1 and AlkG1), which cannot replace the C-terminal domain of AlkG (AlkG2) in alkane oxidation (Kok et al., 1989a ), are almost as highly conserved as AlkG2. This suggests that these domains do play a role in alkane oxidation, perhaps under conditions that are different from those used in the complementation experiments.

In P1, a DNA segment which shows 87% sequence identity to the first half of the GPo1 alkN gene is truncated by the insertion sequence ISPpu4.1 (see below). The gene remnant contains three frameshift mutations relative to the GPo1 alkN sequence and an insert composed of a perfect sevenfold AAATGT repeat, while the ATG-start codon has changed to ATC. The alkN gene remnant is located 135 bp downstream of alkL, which is only slightly more than the distance between the coding regions of the alkBFGHJKL operon, suggesting that in P1 alkN was part of the alkBFGHJKL operon. The gene organization in GPo1 suggests that the IS element ISPpu1 has interrupted a previously existing alkBFGHJKLN operon, leaving the alkN gene without a promoter. In a later evolutionary step, the alkN gene acquired an alkB promoter sequence, which also includes 76 bases of the alkB coding sequence. Interestingly, this sequence has significantly higher sequence identity to the P1 alkB promoter (83% identity) than to the GPo1 alkB promoter (56% identity), suggesting that the source of this DNA may be a different alk operon, more closely related to the P1 alk genes than to the GPo1 alk genes.

Analysis of the alkB, alkS and alkN promoter regions
Apart from the alk gene-coding regions, only the alkB and alkS promoter regions show clear homology. An alignment of the alkB promoter regions of GPo1 and P1 (Fig. 3a) shows that 154 bases upstream of the alkB start codon are conserved with 56% identity. The homology ends with a conserved 20 bp imperfect inverted repeat (IR-B), directly upstream of the -35 region of the alkB promoter (alkBp) characterized previously (Canosa et al., 2000 ; Kok et al., 1989b ). As discussed above, the 230 bp DNA segment upstream of the GPo1 alkN gene is more closely related to the P1 alkB promoter than to the GPo1 alkB promoter. In addition, the homology between the P1 alkB and GPo1 alkN promoter regions extends 15 bases upstream of the inverted repeat (Fig. 3a).

The 103 bases upstream of the alkS ATG start codon in GPo1 and P1 are well conserved at 93% sequence identity, including by 3 bases the {sigma}S-dependent alkS promoter (alkSp1) (Canosa et al., 1999 ). Here, an 18 bp perfect inverted repeat (IR-S), similar to IR-B, is positioned directly upstream of the -35 region of alkSp2, the AlkS-dependent promoter located downstream of alkSp1 (Canosa et al., 2000 ). IR-S overlaps with the -10 region of alkSp1 (Fig. 3b). Based on their positions relative to the promoters, IR-B and IR-S might be control site locations that bind AlkS (Canosa et al., 1999 , 2000 ; Collado-Vides et al., 1991 ; Kok et al., 1989b ). If the inverted repeats bind AlkS, the presence of these sequences in trans on a high-copy-number plasmid should influence expression of a reporter gene controlled by alkBp. To test this hypothesis, we used E. coli JM101 containing a single copy of a cassette composed of xylE under control of alkBp, and alkS* (NdeI site removed), integrated in the chromosome (Panke et al., 1999a ). Three versions of the inverted repeats (Fig. 3c), the first two identical to IR-B and IR-S, and a third in which IR-B and IR-S are combined to a perfect 20 bp inverted repeat (IR-syn), were cloned in the high-copy-number vector pGEM-7Zf(+) and transferred to the reporter strain. As control plasmids pGEM-7Zf(+) and pBG201, a pGEM-7Zf(+)-derived vector containing the alkB gene and the alkBp promoter, were used. The recombinants were cultured in minimal medium with glucose as carbon source to repress the lac promoter of pGEM-7Zf(+), as transcriptional activity of this promoter might replace AlkS from the IR sequences. The recombinant strains were assayed for XylE activity 2 h after induction with the non-metabolizable inducer dicyclopropylketone (Grund et al., 1975 ).

The presence of pGEM-7Zf(+) did not affect XylE activity. Recombinants containing IR-B, IR-S or IR-syn, however, show a strongly reduced XylE activity (to 7–8% for IR-B, 5–6% for IR-S, and 1% for IR-syn), which suggests that AlkS indeed binds to the inverted repeats. Interestingly, pBG201 reduces the XylE activity by only 30–40%, perhaps because transcriptional activity of alkBp in this construct replaces AlkS from its binding site. The competition experiments confirm that AlkS binds to the alkS as well as the alkB promoter (Canosa et al., 2000 ), at positions that are appropriate for control site locations. The fact that the inverted repeats are well conserved between GPo1 and P1 further supports this notion.

As discussed above, the homology between the GPo1 and P1 alk promoter regions ends almost exactly at the inverted repeat of alkBp or the -35 region of alkSp1. In GPo1, fragments of insertion sequences related to IS1384 begin only 15 and 3 bp from the conserved promoter regions (Fig. 3). This indicates that DNA upstream of the conserved regions is not involved in regulation.

DNA flanking the GPo1 and P1 alk gene clusters
Sequence analysis of DNA segments flanking the GPo1 alk genes reveals a mosaic structure of (incomplete) insertion sequences (listed in Table 2) and ORFs (Fig. 2, Table 1). The 5·5 kb DNA segment (55 mol% G+C) directly upstream of the GPo1 alkB gene, the 1·8 kb segment between alkL and alkN, and shorter segments up- and downstream of alkST are entirely composed of insertion sequences or their remnants, and encode proteins with clear homology to transposases or other proteins involved in DNA transposition events. The remaining DNA segments encode proteins that have no homologues in the databases (ORFs 1, 2, 8, 9 and 15), or only have distant relatives that do not allow the assignment of a function (ORFs 10, 12 and 14). The only exceptions are a complete and an incomplete recA gene upstream of alkS, the latter showing 90% DNA sequence identity with P. aeruginosa PAO1 recA.


View this table:
[in this window]
[in a new window]
 
Table 2. Complete and partial insertion sequences flanking the alk gene clusters

 
DNA flanking the P1 alk genes is almost exclusively composed of insertion sequences (6·2 out of 7·4 kb) with a G+C content close to 56 mol% (P1 alk genes: 45%). Only the far end of the EcoRI–SalI fragment containing P1 alkLN' and ISPpu4.1 is not related to insertion sequence elements, but is highly homologous to genes in the P. putida G7 naphthalene lower pathway (Grimm & Harwood, 1999 ).

Based on three criteria (encoded transposase, terminal inverted repeats and flanking direct repeats), six full-length insertion sequences were assigned registration numbers (http://www-is.biotoul.fr/is/is_about.html, 18 October 2000) (Table 2). ISPpu1 is located immediately downstream of alkL in GPo1, and is related to the Agrobacterium tumefaciens and Agrobacterium vitis IS870 copies (Fournier et al., 1993 ). The IS element includes terminal inverted repeats that are quite similar to the IS870 inverted repeats, and is bordered by a 4 bp direct repeat (CTAA). ISPpu2 and ISPpu3 are located between alkS and alkB. They are quite similar to each other (70% DNA sequence identity) and more distantly related to GPo1 ISPpu1 and IS870. In particular, the right terminal inverted repeats (adjacent to each other) of these two IS elements are not well conserved, whilst flanking direct repeats were not present. The two left (outside) inverted repeats are more conserved and ISPpu2 and ISPpu3 together are flanked by the direct repeat TATA, which suggests that they transposed as a single IS element (Table 2). ISPpu5 is located downstream of alkT and is related to the Yersinia enterocolitica IS1328 (64% DNA sequence identity) (Rakin & Heesemann, 1995 ). Alignment of the terminal inverted repeats of ISPpu5 with those of IS1328 shows that the last 6 bases of the left inverted repeat are replaced by the left end of ISPpu4.2, suggesting that ISPpu4.2 inserted itself in the left end of ISPpu5, which removed the direct repeat in the process.

ISPpu4.1 and ISPpu4.2 are almost identical IS elements that show 62% DNA sequence identity to IS1240 (Hanekamp et al., 1997 ) and are located on either side of the P1 alk genes. As discussed above, ISPpu4.1 has truncated the P1 alkN sequence, while ISPpu4.2 is located downstream of alkT. Analysis of the left and right ends of ISPpu4.1, ISPpu4.2 and IS1240 shows that the homology between the left and right ends extends beyond the imperfect inverted repeats (Table 2), which suggests that these extensions are part of the insertion sequence. ISPpu4.1 and ISPpu4.2 individually are not flanked by direct repeats. However, the combination of both insertion sequences (including the alk genes) is flanked by a 4 bp direct repeat (CGTA). In addition, further analysis showed that the entire cassette has interrupted an IS element related to IS401. Although the TnpA ORFs in both ISPpu4 copies are disrupted by frameshift mutations, including a 23 bp deletion in ISPpu4.2, which most likely has inactivated both insertion sequences, the above features are evidence for a class I transposon (Kleckner, 1981 ), which we have named TnPpu-alk1.

Analysis of the GPo1 sequence did not reveal a similar structure. However, a 638 bp DNA segment starting only 6 bases upstream of the alkS promoter has 77% sequence identity to the right end of IS1384 (accession no. AF052751), whilst an 88 bp region ending only 15 bp from the alkB promoter shows 86% identity to an internal fragment of the same insertion sequence (the ends of these DNA segments are marked in Fig. 3). These fragments do not overlap, which makes it impossible to ascertain whether the IS1384-related sequences were identical. Nevertheless, the arrangement suggests that the GPo1 alk genes were part of a class I transposon as well, before large segments of the flanking insertion sequences were lost. This analysis also demonstrates that class I transposons are not very stable; they are formed by the fortuitous insertion of the same IS element on both sides of a given set of genes, and become inactive again when one of the constituting IS elements is destroyed, for example by the insertion of other IS elements, deletions or frameshift mutations. Eventually, mosaics of (incomplete) insertion sequences flank the catabolic genes. Such sections are, as they no longer contain functional genes or genes necessary for the survival of the host organism, selected targets for new transposition events, which destroy previously inserted IS elements, but also provide new opportunities for the mobilization of catabolic genes, such as the alk gene clusters. Analysis of the alk genes of two P. putida strains provides a clear example of this evolutionary process in action and at the same time has helped us to identify new functions and aspects of alkane degradation.


   ACKNOWLEDGEMENTS
 
This research was supported by the Swiss National Science Foundation through the Swiss Priority Program in Biotechnology, grant no. 5002-037023.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990). Basic local alignment search tool. J Mol Biol 215, 403-410.[Medline]

Bachmann, B. (1987). Derivations and genotypes of some mutant derivatives of Escherichia coli K-12. In Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, pp. 1190–1219. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.

Baptist, J. N., Gholson, R. K. & Coon, M. J. (1963). Hydrocarbon oxidation by a bacterial enzyme system: I. Products of octane oxidation. Biochim Biophys Acta 69, 40-47.

van Beilen, J. B. (1994). Alkane oxidation by Pseudomonas oleovorans: genes and proteins. PhD thesis, University of Groningen.

van Beilen, J. B., Eggink, G., Enequist, H., Bos, R. & Witholt, B. (1992a). DNA sequence determination and functional characterization of the OCT-plasmid-encoded alkJKL genes of Pseudomonas oleovorans. Mol Microbiol 6, 3121-3136.[Medline]

van Beilen, J. B., Penninga, D. & Witholt, B. (1992b). Topology of the membrane-bound alkane hydroxylase of Pseudomonas oleovorans. J Biol Chem 267, 9194-9201.[Abstract/Free Full Text]

van Beilen, J. B., Wubbolts, M. G. & Witholt, B. (1994). Genetics of alkane oxidation by Pseudomonas oleovorans. Biodegradation 5, 161-174.[Medline]

Benson, S., Fennewald, M., Shapiro, J. & Huettner, C. (1977). Fractionation of inducible alkane hydroxylase activity in Pseudomonas putida and characterization of hydroxylase-negative plasmid mutations. J Bacteriol 132, 614-621.[Medline]

Bustamante, C., Gurrieri, S. & Smith, S. B. (1993). Towards a molecular description of pulsed-field gel electrophoresis. Trends Biotechnol 11, 23-30.[Medline]

Canosa, I., Yuste, L. & Rojo, F. (1999). Role of the alternative sigma factor {sigma}s in the expression of the AlkS regulator of the Pseudomonas oleovorans alkane degradation pathway. J Bacteriol 181, 1748-1754.[Abstract/Free Full Text]

Canosa, I., Sanchez-Romero, J. M., Yuste, L. & Rojo, F. (2000). A positive feedback mechanism controls expression of AlkS, the transcriptional regulator of the Pseudomonas oleovorans alkane degradation pathway. Mol Microbiol 35, 791-799.[Medline]

Chakrabarty, A. M., Chou, G. & Gunsalus, I. C. (1973). Genetic regulation of octane dissimulation plasmid in Pseudomonas. Proc Natl Acad Sci USA 70, 1137-1140.[Abstract]

Collado-Vides, J., Magasanik, B. & Gralla, J. D. (1991). Control site location and transcriptional regulation in Escherichia coli. Microbiol Rev 55, 371-394.

Dower, W. J., Miller, J. F. & Ragsdale, C. W. (1988). High efficiency transformation of E. coli by high voltage electroporation. Nucleic Acids Res 16, 6127.[Abstract]

Eggink, G., Lageveen, R. G., Altenburg, B. & Witholt, B. (1987a). Controlled and functional expression of Pseudomonas oleovorans alkane utilizing system in Pseudomonas putida and Escherichia coli. J Biol Chem 262, 17712-17718.[Abstract/Free Full Text]

Eggink, G., van Lelyveld, P. H., Arnberg, A., Arfman, N., Witteveen, C. & Witholt, B. (1987b). Structure of the Pseudomonas putida alkBAC operon. Identification of transcription and translation products. J Biol Chem 262, 6400-6406.[Abstract/Free Full Text]

Eggink, G., Engel, H., Vriend, G., Terpstra, P. & Witholt, B. (1990). Rubredoxin reductase of Pseudomonas oleovorans. Structural relationship to other flavoprotein oxidoreductases based on one NAD and two FAD fingerprints. J Mol Biol 212, 135-142.[Medline]

Fennewald, M. & Shapiro, J. (1977). Regulatory mutations of the Pseudomonas plasmid alk regulon. J Bacteriol 132, 622-627.[Medline]

Fennewald, M., Prevatt, W., Meyer, R. & Shapiro, J. (1978). Isolation of Inc P-2 plasmid DNA from Pseudomonas aeruginosa. Plasmid 1, 164-173.[Medline]

Fennewald, M., Benson, S., Oppici, M. & Shapiro, J. (1979). Insertion element analysis and mapping of the Pseudomonas plasmid alk regulon. J Bacteriol 139, 940-952.[Medline]

Fournier, P., Paulus, F. & Otten, L. (1993). IS870 requires a 5'-CTAG-3' target sequence to generate the stop-codon for its large ORF1. J Bacteriol 175, 3151-3160.[Abstract]

Grimm, A. C. & Harwood, C. S. (1999). NahY, a catabolic plasmid-encoded receptor required for chemotaxis of Pseudomonas putida to the aromatic hydrocarbon naphthalene. J Bacteriol 181, 3310-3316.[Abstract/Free Full Text]

Grund, A., Shapiro, J., Fennewald, M., Bacha, P., Leahy, J., Markbreiter, K., Nieder, M. & Toepfer, M. (1975). Regulation of alkane oxidation in Pseudomonas putida. J Bacteriol 123, 546-556.[Medline]

Hanekamp, T., Kobayashi, D., Hayes, S. & Stayton, M. M. (1997). Avirulence gene D of Pseudomonas syringae pv. tomato may have undergone horizontal gene transfer. FEBS Lett 415, 40-44.[Medline]

Hauben, L., Vauterin, L., Swings, J. & Moore, E. R. B. (1997). Comparison of 16S ribosomal DNA sequences of all Xanthomonas species. Int J Syst Bacteriol 47, 328-335.[Abstract/Free Full Text]

Henikoff, S. (1984). Unidirectional digestion with exonuclease-III creates targeted breakpoints for DNA sequencing. Gene 28, 351-359.[Medline]

Innis, M. A., Gelfand, D. H., Sninsky, J. J. & White, T. J. (1990). PCR Protocols. A Guide to Methods and Applications. San Diego: Academic Press.

Kado, C. I. & Liu, S.-T. (1981). Rapid procedure for detection and isolation of large and small plasmids. J Bacteriol 145, 1365-1373.[Medline]

Kleckner, N. (1981). Transposable elements in prokaryotes. Annu Rev Genet 15, 341-404.[Medline]

Kok, M., Oldenhuis, R., van der Linden, M. P. G., Meulenberg, C. H. C., Kingma, J. & Witholt, B. (1989a). The Pseudomonas oleovorans alkBAC operon encodes two structurally related rubredoxins and an aldehyde dehydrogenase. J Biol Chem 264, 5442-5451.[Abstract/Free Full Text]

Kok, M., Oldenhuis, R., van der Linden, M. P. G., Raatjes, P., Kingma, J., van Lelyveld, P. H. & Witholt, B. (1989b). The Pseudomonas oleovorans alkane hydroxylase gene. Sequence and expression. J Biol Chem 264, 5435-5441.[Abstract/Free Full Text]

Lageveen, R. G., Huisman, G. W., Preusting, H., Ketelaar, P. E. F., Eggink, G. & Witholt, B. (1988). Formation of polyester by Pseudomonas oleovorans: the effect of substrate on the formation and composition of poly-(R)-3-hydroxyalkanoates and poly-(R)-3-hydroxyalkenoates. Appl Environ Microbiol 54, 2924-2932.

Lee, M. & Chandler, A. C. (1941). A study of the nature, growth and control of bacteria in cutting compounds. J Bacteriol 41, 373-386.

Luria, S. E., Adam, J. N. & Teng, R. C. (1960). Transduction of lactose utilizing ability among strains of Escherichia coli and Shigella dysenteriae and the properties of the transducing phage particle. Virology 12, 348-390.[Medline]

Mahillon, J. & Chandler, M. (1998). Insertion sequences. Microbiol Mol Biol Rev 62, 725-774.[Abstract/Free Full Text]

Nozaki, M. (1970). Metapyrocatechase. Methods Enzymol 17, 522-525.

Owen, D. J. (1986). Molecular cloning and characterization of sequences from the regulatory cluster of the Pseudomonas plasmid alk system. Mol Gen Genet 203, 64-72.[Medline]

Panke, S., de Lorenzo, V., Kaiser, A., Witholt, B. & Wubbolts, M. G. (1999a). Engineering of a stable whole-cell biocatalyst capable of (S)-styrene oxide formation for continuous two-liquid phase applications. Appl Environ Microbiol 65, 5619-5623.[Abstract/Free Full Text]

Panke, S., Meyer, A., Huber, C. M., Witholt, B. & Wubbolts, M. G. (1999b). An alkane-responsive expression system for the production of fine chemicals. Appl Environ Microbiol 65, 2324-2332.[Abstract/Free Full Text]

Rakin, A. & Heesemann, J. (1995). Virulence-associated fyuA/irp2 gene cluster of Yersinia enterocolitica biotype 1B carries a novel insertion sequence IS1328. FEMS Microbiol Lett 129, 287-292.[Medline]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Schwartz, R. D. (1973). Octene epoxidation by a cold-stable alkane-oxidizing isolate of Pseudomonas oleovorans. Appl Microbiol 25, 574-577.[Medline]

Schwartz, R. D. & McCoy, C. J. (1973). Pseudomonas oleovorans hydroxylation-epoxidation system: additional strain improvements. Appl Microbiol 26, 217-218.[Medline]

Smits, T. H. M., Röthlisberger, M., Witholt, B. & van Beilen, J. B. (1999). Molecular screening for alkane hydroxylase genes in Gram-negative and Gram-positive strains. Environ Microbiol 1, 307-318.[Medline]

Stanier, R. Y., Palleroni, N. J. & Doudoroff, M. (1966). The aerobic pseudomonads: a taxonomic study. J Gen Microbiol 43, 159-271.[Medline]

Taguchi, K., Fukutomi, H., Kuroda, A., Kato, J. & Ohtake, H. (1997). Genetic identification of chemotactic transducers for amino acids in Pseudomonas aeruginosa. Microbiology 143, 3223-3229.[Abstract]

Tan, H.-M. (1999). Bacterial catabolic transposons. Appl Microbiol Biotechnol 51, 1-12.[Medline]

Witholt, B., de Smet, M. J., Kingma, J., van Beilen, J. B., Kok, M., Lageveen, R. G. & Eggink, G. (1990). Bioconversions of aliphatic compounds by Pseudomonas oleovorans in multiphase bioreactors: background and economic potential. Trends Biotechnol 8, 46-52.[Medline]

Received 25 October 2000; revised 2 February 2001; accepted 13 February 2001.