OUP user menu

Identification and characterization of Pseudomonas aeruginosa PA-IIL lectin gene and protein compared to PA-IL

Nechama Gilboa-Garber, Don J. Katcoff, Nachman C. Garber
DOI: http://dx.doi.org/10.1111/j.1574-695X.2000.tb01505.x 53-57 First published online: 1 September 2000

Abstract

Using the 33 N-terminal amino acids of the fucose/mannose binding lectin PA-IIL of Pseudomonas aeruginosa ATCC 33347 in a tblastn search of P. aeruginosa PAO1 genomic sequence in GenBank revealed a single open reading frame encoding a 114-amino acid protein (excluding initiator methionine) perfectly matching that amino acid sequence. Following its stop codon there is a GC-rich sequence having a perfect dyad symmetry promoting formation of a hairpin loop structure, potentially enabling rho-independent transcription termination. Upstream of the putative ribosomal binding site there are sequences resembling Vibrio fischeri luxI box, consistent with autoinduction of this gene. The predicted PA-IIL molecular mass, confirmed by mass spectrometry, is 11 732 Da. Its pI is 3.88. The C-terminal domain is particularly hydrophobic, implying possible embedding in the cell membrane. PA-IIL is similar to P. aeruginosa PA-IL lectin in some amino acids and potential glycosylation sites but lacks cysteine, methionine and histidine. Despite their relations in functions and regulation, their genes are widely separated (by about 867.5 kb).

Keywords
  • Pseudomonas aeruginosa
  • Lectin gene
  • Virulence factor
  • Lectin structure

1 Introduction

The bacterium Pseudomonas aeruginosa produces two potent lectins, PA-IL and PA-IIL, which are associated and coexpressed with its cytotoxic virulence factors [1]. The first lectin is galactophilic and the second binds fucose and mannose. Both lectins were purified by affinity chromatography and their properties, mode of regulation, diverse biological effects and some important applications for science and medicine were studied and described [13]. The PA-IL gene (pa-1L) was isolated from a P. aeruginosa ATCC 27853 genomic library [4] and expressed in Escherichia coli [5]. It contains a 369-bp ORF (open reading frame). Its 121 amino acid sequence, deduced from the DNA sequence, predicted a molecular mass of 12 763 Da fitting well with the experimentally measured molecular mass and with amino acid analyses [4,5]. In the present communication, we definitively identify the pa-2L gene and deduced PA-IIL 114-amino acid sequences. To locate the gene ORF, we used the 33 N-terminal amino acid sequence, determined experimentally, to perform a tblastn query in the P. aeruginosa PAO1 sequence database containing the almost complete genome [6]. The identity of the resulting ORF which perfectly matched the sequence used in the query was confirmed by comparing the predicted molecular mass to the precise molecular mass of PA-IIL as determined by mass spectrometry. The identification of the pa-2L gene and the deduced protein allows a better understanding of PA-IIL properties and the coordinate expression of both lectins, which contribute to the pathogenicity of the bacterium [2,3].

2 Materials and methods

2.1 Lectin amino acid sequencing

The two lectins PA-IL and PA-IIL were purified from stationary phase P. aeruginosa ATCC 33347 cell extracts using affinity chromatography [1] and their N-terminal amino acid sequence was determined by Edman degradation performed as previously described [4].

2.2 Tblastn search

As the sequencing of the P. aeruginosa genome has been almost completed, we performed a tblastn search of the P. aeruginosa PAO1 sequence deposited at the Pseudomonas Genome Project (PGP) (http://www.pseudomonas.com) updated on December 15, 1999, using the N-terminal 33-amino acid sequence of PA-IIL purified from P. aeruginosa ATCC 33347.

2.3 Genetic analysis

The resulting ORF and surrounding sequences were analyzed by the Genetics Computer Group package of software installed under a Digital UNIX V4.0C operating system.

2.4 Mass spectrometry analyses

The purified lectin preparations (giving a single band in gel electrophoresis) [1] were used. Mass spectrometry was performed by the Technion Protein Research Center using a matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometer (MALDI-TOF 2E, Micromass, Manchester, UK).

3 Results and discussion

A single ORF encoding a 114-amino acid protein (excluding initiator methionine) (Fig. 1) was found that perfectly matched the query sequence at its N-terminus. To confirm the identification of this ORF as belonging to PA-IIL, the precise molecular mass of purified PA-IIL was measured by mass spectrometry (Fig. 2) performed in the Technion Protein Research Center (Haifa). This technique yielded a molecular mass of 11 732.0 Da. The calculated molecular mass of the putative protein deduced from the ORF without the initiator methionine was 11 731.8 Da when isotopic frequency was taken into account. Thus, the identification of this ORF belonging to PA-IIL is absolute and we can definitively say that the vast majority, if not all, of the PA-IIL in the cell does not contain the initiator methionine, and is not glycosylated. Since PA-IIL shares many properties and is coordinately expressed with PA-IL [7], we compared their DNA and amino acid sequences (Fig. 1) and performed MALDI-TOF mass spectrometry on PA-IL as well (Fig. 2). Fig. 2 shows that the experimental molecular mass of PA-IL was found to be 12 763.0 Da, which reconfirmed the earlier identification of its gene [4,5]. As can be seen in Fig. 1, PA-IIL is seven amino acids shorter than PA-IL, which contains 121 amino acids, and it has numerous tiny amino acid repeats whose function is unknown. Both lectins have similar alanine, arginine, glutamine and glutamic acid contents and contain potential glycosylation sites situated at 107 in PA-IL and at 22 and 35 in PA-IIL (Fig. 1). PA-IIL lacks cysteine, methionine and histidine, has more valine, asparagine, threonine and serine but fewer glycine and isoleucine residues than PA-IL. The predicted pI of PA-IIL is more acid, being 3.88 compared to 4.94 for PA-IL. Predicted secondary structures of PA-IL and PA-IIL (Chou—Fasman prediction), as generated by the PLOTSTRUCTURE program in the GCG package, including hydrophobic and hydrophilic regions, as well as potential glycosylation sites, showed that unlike PA-IL, PA-IIL has a C-terminal domain that is particularly hydrophobic, implying that it may be embedded in the cell membrane.

Figure 1

Nucleotide and deduced amino acid sequences of PA-IL and PA-IIL genes and proteins. The numbers of nucleotides are indicated on the left and right sides. The nucleotide sequences of PA-IL and PA-IIL genes are as presented in the contig described in the Pseudomonas Genome Project database at locations 2 905 499–2 905 134 (pa-1L) and 3 772 978–3 773 322 (pa-2L). The N-terminal 33-amino acid sequence of PA-IIL used for the query is identical to the deduced one without the initiator methionine. The potential N-glycosylation sites of PA-IL (at 107) and PA-IIL (at 22 and 35) are labeled by a square on the respective amino acids. The initiation and stop codons of both lectins are printed in bold letters. Dashes indicate luxI box-like sequences and the Shine—Dalgarno sequences in both are located 9 bp upstream of the start codon. The underline indicates a dyad symmetry.

Figure 2

Determination of the precise molecular masses of purified PA-IL (A) and PA-IIL (B) using MALDI-TOF mass spectrometry.

About 16 bp following the stop codon of PA-IIL there is a GC-rich 24-bp sequence having perfect dyad symmetry (Fig. 1). This sequence may form a hairpin loop structure, potentially enabling rho-independent transcription termination. In the PA-IL gene, there are two consecutive stop codons, followed by a potential hairpin loop with seven consecutive A and respective T residues [4,5]. Nine base pairs upstream of the translational start site of both lectins there is a putative Shine—Dalgarno sequence (ribosomal binding site). Further upstream of pa-2L there are three sequences with a certain homology to the lux box originally described in Vibrio fischeri [8] and later also in other P. aeruginosa PAO1 genes including pa-1L [9] as described by P. Williams et al. (personal communication) (Figs. 1 and 3). As PA-IIL expression can be induced by acylhomoserine lactone autoinducers as described for several P. aeruginosa virulence genes [1013] including those of the lectins [9,14], it is possible that one or more of these sequences may also be involved in the PA-IIL autoinduction process. In spite of the close relationship between the two lectins (Table 1), their genes are widely (about 867.5 kb) separated on the P. aeruginosa chromosome. They are aligned on the single contig in the PGP database (December, 15, 1999) where the pa-1L ORF is situated between bp 2 905 499 and 2 905 134 while the pa-2L ORF is located between bp 3 772 978 and 3 773 322 according to the numbering scheme in that database. Therefore it seems that their coordinate expression may be only attributed to the common autoinduction process via the lux-like boxes.

Figure 3

Nucleotide sequences in the upstream region of the PA-IIL gene (pa-2L) resembling a lux box as compared to that of pa-1L, other genes of P. aeruginosa and the luxI box of V. fischeri [9]. Identical residues are shown with a dark background.

View this table:
Table 1

Partial comparison of PA-IL and PA-IIL properties

PA-ILPA-IIL
Sugar specificitymelibiose>methyl-α-Gal>Gal>methyl-β-Gal>GalNAcp-nitrophenyl-α-l-fucose>l-fucose>fucosylamine=l-Gal>d-mannose>d-fructose
Number of amino acids (excluding initiator methionine)121114
Molecular mass (Da)12 76311 732
pI4.943.88
C-terminal domainhydrophilichydrophobic
Methionine in the N-terminusabsentabsent
Cysteine, methionine and histidine presencepresentabsent
Cations in the moleculeCa2+ Mg2+Zn2+ Ca2+ Mg2+[3]
Potential glycosylation sitepresentpresent
Gene location on the chromosome (contig in PGP)bp 2 905 499–2 905 134bp 3 772 978–3 773 322
Ribosomal binding site (Shine—Dalgarno sequence)located 9 bp upstream of the translation start sitelocated 9 bp upstream of the translation start site

Acknowledgements

The research was supported in part by the Health Sciences Center of the Faculty of Life Sciences, and by research fund of Bar-Ilan University. The authors wish to thank Mrs. Avrille Goldreich and Mrs. Ella Gindi for the useful help in the preparation of the manuscript and the graphic presentation.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
View Abstract