Archives of Microbiology

, Volume 190, Issue 4, pp 439–449

Identification of heme uptake genes in the fish pathogen Aeromonas salmonicida subsp. salmonicida

Authors

  • Mohsen Najimi
    • Department of Microbiology and Parasitology, Institute of Aquaculture and Faculty of BiologyUniversity of Santiago de Compostela
  • Manuel L. Lemos
    • Department of Microbiology and Parasitology, Institute of Aquaculture and Faculty of BiologyUniversity of Santiago de Compostela
    • Department of Microbiology and Parasitology, Institute of Aquaculture and Faculty of BiologyUniversity of Santiago de Compostela
Original Paper

DOI: 10.1007/s00203-008-0391-5

Cite this article as:
Najimi, M., Lemos, M.L. & Osorio, C.R. Arch Microbiol (2008) 190: 439. doi:10.1007/s00203-008-0391-5

Abstract

Aeromonas salmonicida subsp. salmonicida, the causative agent of furunculosis in fish, can use heme as the sole iron source. We applied the Fur Titration Assay to isolate a cluster including six genes hutAZXBCD that showed similarity to heme uptake genes of other Gram-negative bacteria, and three genes orf123 of unknown function. The spatial organization of these nine genes, arranged in five transcriptional units, was similar to that of a homologous cluster in A. hydrophila. When a TonB system was provided, this cluster allowed Escherichia coli 101ESD (an ent mutant, unable to synthesize enterobactin) to utilize hemin and hemoglobin as iron sources. Mutation of hutB, a gene that encodes a predicted periplasmic hemin-binding protein, caused a drastic defect in the ability of A. salmonicida to grow with hemin as unique source of iron. Interestingly, a mutant for hutA gene (encoding the outer membrane hemin receptor) showed initially a reduced ability to grow with hemin as sole iron source, but after 24 h it achieved growth levels similar to parental strain. Thus mutation of hutA could not abolish the growth with hemin as iron source, suggesting that redundant outer membrane heme transport functions might be encoded in the A. salmonicida genome.

Keywords

Aeromonas salmonicidaIron uptakeHeme uptakeHemin

Introduction

Bacterial pathogens have developed efficient mechanisms to obtain iron from the host, due to the lack of readily available iron in biological systems (Ratledge and Dover 2000). One of the main strategies is the synthesis and secretion of siderophores, which can remove iron from host iron-binding proteins (Crosa and Walsh 2002). Although siderophores bind iron with high affinity, they cannot remove iron from heme. Many Gram-negative pathogens have the ability to obtain iron from free heme or heme proteins by means of siderophore-independent mechanisms (Genco and Dixon 2001; Osorio and Lemos 2002). Expression of most genes required for obtaining iron, including siderophore biosynthesis and transport genes as well as heme uptake genes, are regulated by the iron-binding repressor protein Fur (ferric uptake regulator) (Braun et al. 1998).

Aeromonas salmonicida subsp. salmonicida is the causative agent of the fish disease known as furunculosis, a cause of significant economic losses in salmonid and turbot aquaculture worldwide (Toranzo et al. 2005). A. salmonicida subsp. salmonicida is known to produce siderophores for iron uptake (Chart and Trust 1983; Hirst et al. 1991; Fernandez et al. 1998), and a gene cluster encoding determinants for the biosynthesis of a catechol-type siderophore has been described recently (Najimi et al. 2008). Although it is known that certain atypical A. salmonicida strains require heme as a porphyrin source for growth in laboratory media (Ishiguro et al. 1986), to date it has not been reported whether heme compounds can be used as sole iron sources by this fish pathogen, and no heme uptake genes have been characterized at the functional level in A. salmonicida. Using a proteomic approach, a recent study demonstrated that under iron-limited conditions A. salmonicida expresses three iron-regulated outer membrane receptors, and one of these receptors was proposed to be a putative heme receptor based on sequence homology (Ebanks et al. 2004).

Heme utilization genes have been identified in numerous species, including fish pathogens as Photobacterium damselae (Rio et al. 2005), Vibrio anguillarum (Mazoy et al. 2003; Mouriño et al. 2004) and important human pathogenic bacteria as Yersinia enterocolitica (Stojiljkovic and Hantke 1992), V. cholerae (Henderson and Payne 1993; Occhino et al. 1998; Mey and Payne 2001) and Shigella dysenteriae (Mills and Payne 1997). The involvement of heme uptake in the virulence of bacterial pathogens has also been demonstrated in several species, such as V. cholerae (Henderson and Payne 1994), and Neisseria meningitidis (Stojiljkovic et al. 1995). Heme-uptake in gram negative bacteria usually involve outer membrane receptors as well as a TonB-dependent internalization process with two accessory proteins ExbB and ExbD, and this system is believed to transduce the energy of the proton motive force of the cytoplasmic membrane into transport energy required by the receptor. Subsequently, transport of heme across the cytoplasmic membrane is driven by ATP hydrolysis, and an ATP-binding cassette (ABC) transporter is involved in this transport (Köster 2001).

The goal of the present study was to identify genetic determinants involved in heme uptake in A. salmonicida subsp. salmonicida. An heme uptake cluster was sequenced and analyzed, and deletion mutants in two different genes were constructed and tested for their ability to grow with hemin as the unique source of iron. Complementation of Escherichia coli mutants with the A. salmonicida genes was performed in order to evaluate the restoration of heme utilization as an iron source.

Materials and methods

Bacterial strains, plasmids and culture conditions

Bacterial strains and plasmids used in this study are listed in Table 1. A. salmonicida subsp. salmonicida strains were routinely grown at 22°C in tryptic soy agar and broth (TSB) (Difco) supplemented with 1% NaCl as well as in M9 minimal medium supplemented with 0.2% Casamino Acids (Difco) (CM9). E. coli strains were grown at 37°C in Luria-Bertani medium or CM9 supplemented with antibiotics when required. Antibiotics were used at the following final concentrations: kanamycin (Km) at 25 μg ml−1, ampicillin (Ap) sodium salt at 50 μg ml−1, chloramphenicol (Cm) at 20 μg ml−1 and tetracycline (Tet) hydrochloride at 15 μg ml−1. All stocks were filter sterilized and stored at −20°C. CM9 medium was made low in iron by the addition of ethylenediamine di (o-hydroxy-phenylacetic acid) (EDDA) (Sigma-Aldrich) or 2,2’-dipyridyl at different concentrations, as chelators of non-heme iron. Stock solution of EDDA was prepared solving 1 g in 1 N NaOH, adjusting to pH 9.0 with 6 N HCl and bringing volume up to 20 ml in ultrapure water. The concentration of this solution is 138 mM and was assayed at different final concentrations. 2,2’-dipyridyl was prepared at 10 mM in ultrapure water. Bovine hemin (Sigma) was dissolved at 5 mM in 10 mM NaOH. Bovine hemoglobin (Sigma) was dissolved at 1 mM in water.
Table 1

Bacterial strains and plasmids used in this study

Strain or plasmid

Relevant characteristics

Source or reference

A. salmonicida

  

ACR168.1

 

Turbot, Spain

RSP74.1

CmR

Turbot, Portugal

MON15

RSP74.1 siderophore biosynthesis mutant (∆asbD)

Najimi et al. (2008)

MON29

MON15 hutA defective mutant

This study

MON30

MON15 hutB defective mutant

This study

E. coli

  

DH5α

Cloning strain

Laboratory stock

S17-1-λpir

recAthiprohsdR-M+ RP4-2-Tc:Mu-Km:Tn7 λpir

Herrero et al. (1990)

H1717

araD139 ΔlacU169rpsL150 relA1 flb5301 deoC1ptsF25rbsR aroB fhuF::λ placMu

Hantke (1987)

XL1-Blue MR

Δ(mcrA)183Δ(mcrCB-hsdSMR-mrr)173 endA1supE44thi-1recA1gyrA96 relA1 lac

Stratagene

101ESD

HB101 derivative, deficient in enterobactin biosynthesis Δ(entC-entA)

Crosa and Walsh (2002)

Plasmids

  

pWKS30

Low-copy-number cloning vector, ApR

Wang and Kushner (1991)

pACYC184

Cloning vector, TetR, CmR

Rose (1988)

pKEK229

Suicide vector derived from pCVD442, pir, ApR

Correa et al. (2000)

PT7-7

Cloning vector, ApR

Tabor and Richardson (1985)

SuperCos1

Cosmid vector, ApR

Stratagene

pFMON12

FURTA-positive clone of A. salmonicida ACR168.1 containing partial heme receptor hutA and promoter region cloned into PT7-7, ApR

This study

pFMON42

FURTA-positive clone of A. salmonicida ACR168.1 containing hutZ, hutX, hutB and promoter regions cloned into PT7-7, ApR

This study

pGMON5

Cosmid clone from A. salmonicida ACR168.1 containing hutA-orf1-orf2-orf3-hutZX-hutBCD genes, ApR

This study

pML1

Cosmid clone containing the heme uptake cluster of Vibrioanguillarum

Mazoy et al. (2003)

pCAR168

tonB-exbBD genes of Photobacterium damselae subsp. damselae cloned into pGEMT-Easy, ApR

This study

pSML33

tonB-exbBD and huvBCD genes of V. anguillarum cloned into pWKS30, ApR

Mouriño et al. (2004)

pMON33

hutA-orf1-orf2-orf3-hutZX-hutBCDgenes from A. salmonicida ACR168.1 cloned into pACYC184, TetR

This study

pMON34

hutA from A. salmonicida ACR168.1 cloned into pACYC184, TetR

This study

pMON36

hutAorf1 orf2 orf3 hutZX from A. salmonicida ACR168.1 cloned into pACYC184, TetR

This study

Recombinant DNA techniques, DNA sequencing and data analysis

Total genomic DNA was extracted and purified with the Easy-DNA kit (Invitrogen). Plasmid DNA purification and DNA extraction from agarose gels were performed with kits from QIAGEN. DNA sequences were determined by the dideoxy chain termination method using the CEQ DTCS-Quick Start Kit (Beckman Coulter) using a capillary DNA sequencer CEQ8000 (Beckman Coulter). Comparison of the sequence data with published sequences in EMBL/GenBank was performed with the blast software via the Internet (http://ncbi.nlm.nih.gov/BLAST and http://www.ebi.ac.uk/blast/index.html). Prediction of protein domains was carried out by using the Pfam database online facilities (http://www.sanger.ac.uk/Software/Pfam/). Additional DNA and peptide analyses were performed using Bioedit (Hall 1999).

Fur titration assay and screening of a A. salmonicida cosmid library

Fur regulated promoters and iron-binding proteins carried on a multicopy plasmid can be identified by transformation into E. coli strain H1717, that carries a Fur-regulated fhuF::lacZ gene fusion. Fur boxes introduced on a multicopy plasmid can cause derepression of the fusion by titrating the Fur protein, and thus leading to the expression of a Lac+ phenotype. In this study we sequenced ca. 100 clones of a plasmid library of A. salmonicida ACR168.1 constructed in the indicator strain H1717 in a previous work (Najimi et al. 2008), and selected two Fur Titration Assay (FURTA)-positive clones named pFMON12 and pFMON42 that contained genes showing similarity to putative heme uptake genes in other Gram-negative bacteria.

We screened an A. salmonicida ACR168.1 cosmid library (Najimi et al. 2008) by colony-PCR using primers targeted to genes encoded in plasmids pFMON12 and pFMON42. This led to isolation of cosmid pGMON5. Cosmid DNA was purified using the QIA filter Midi Kit (QIAGEN) and used for DNA sequencing.

RNA isolation and RT-PCR

In order to determine the transcriptional organization of the heme gene cluster, A. salmonicida RSP74.1 was grown in TSB-1 and then subcultured on CM9 minimal medium supplemented with 60 μM EDDA. Total RNA was extracted using the RNAwiz Isolation Reagent (Ambion). RNA preparations were treated with RQ1 RNase-free DNaseI (Promega). Reverse transcription reaction was carried out using the M-MLV reverse transcriptase (Invitrogen), and primers RT-1, RT-2, RT-3 and RT-4, targeted to the 3′-end of genes orf3, orf1, hutZ and hutD, respectively (Fig. 1a). For each reaction, 1 μg of total RNA was used. Subsequent PCR amplification was performed by using suitable primer pairs for each tested gene. As positive controls, each primer combination was tested in PCR reactions using chromosomal DNA as a template. A negative control reaction for PCR was performed with total RNA without M-MLV reverse transcriptase to confirm the lack of genomic DNA contamination in each reaction mixture.
https://static-content.springer.com/image/art%3A10.1007%2Fs00203-008-0391-5/MediaObjects/203_2008_391_Fig1_HTML.gif
Fig. 1

a Physical map of the heme uptake cluster of Aeromonas salmonicida. Open reading frames are depicted as white thick arrows which indicate the direction of transcription. A circle with an “F” underneath denotes presence of a predicted Fur box. The white triangle with “Term” on top denotes a potential transcriptional terminator. RT-1, RT-2, RT-3 and RT-4 are the primers used for Reverse-transcription for operon mapping. Thick lines referred to as pFMON12 and pFMON42 denote the position within the gene map above, of the DNA sequence cloned in these two FURTA-positive clones. The three thin lines under the gene map depict the relevant subclones of the heme cluster, pMON33, pMON34 and pMON36. A scale for 1 kb is provided at top right b Operon mapping: results of PCR amplifications using as templates the products of RT reactions obtained with primers RT-1 (for orf2), RT-2 (for orf1), RT-3 (for hutX) and RT-4 (for hutB). M molecular size marker (100 bp). Negative controls (−) are PCR reactions without reverse transcriptase. Positive controls (+) are PCR reactions using chromosomal DNA as template

Construction of a hutA and hutB-targeted in-frame deletion mutants

Deletion mutants of the hutA and hutB genes in A. salmonicida MON15 (Table 1) were constructed by allelic exchange as previously described (Correa et al. 2000; Mouriño et al. 2004). In brief, gene deletion was carried out by using PCR amplifications of two fragments of the gene and flanking regions, which when ligated together would result in an in-frame (non polar) deletion. This process resulted in the formation of mutant alleles ΔhutA (removes coding sequence for amino acids 45-657) and ΔhutB (removes coding sequence for amino acids 107-242). The deleted alleles were cloned in the suicide vector pKEK229 (Correa et al. 2000). As a pCVD442 derivative, pKEK229 contains R6K ori, requiring the pir gene product for replication, and the sacB gene conferring sucrose sensitivity. The resulting plasmids were mated from E. coli S17-1-λpir into A. salmonicida subsp. salmonicida MON15, a derivative of strain RSP74.1 that cannot produce siderophores (Najimi et al. 2008) and transformants with the plasmid integrated in the chromosome by homologous recombination were selected on agar medium containing 50 μg ml−1 Ap (resistance conferred by pKEK229) and 20 μg ml−1 Cm (to select A. salmonicida MON15). A second recombination event was obtained by selecting for sucrose resistance (10% wt/vol) and resistance to chloramphenicol. This led to obtention of A. salmonicida MON29 (ΔhutA) and MON30 (ΔhutB) (Table 1).

Complementation of E. coli 101ESD

We tested whether the A. salmonicida heme uptake gene cluster identified after partial sequencing of cosmid pGMON5 could provide E. coli 101ESD Δ(entC-entA) with all the necessary functions to utilize hemin and hemoglobin as sole iron source. This E. coli strain lacks heme transport genes and thus constitutes an ideal heterologous host to reconstruct a heme uptake mechanism by means of providing gene combinations in recombinant plasmids. Plasmids containing different gene combinations (Fig. 1a; Tables 1, 3) were transformed in E. coli 101ESD. About 100 μl of overnight cultures of each transformant were added to 3 ml of molten soft CM9 medium and plated onto appropriate prepoured CM9 plates supplemented with 110 μM 2,2’-dipyridyl. Sterile filter-paper disks were loaded with 20 μl of either 5 mM hemin or 1 mM hemoglobin. Disks spotted with 5 mM FeSO4 were included as positive controls for utilization of iron sources. Results were annotated as positive or negative after 24 h of incubation.

Growth under iron-limiting conditions and assay methods for hemin utilization

In order to test the ability of A. salmonicida MON15 and the respective hutA and hutB defective mutants to grow in the presence of hemin as sole iron source, overnight CM9 cultures of parental and mutant strains were adjusted to an OD600 of 1, diluted 1:200 in fresh CM9 minimal medium supplemented with the iron chelator EDDA at 60 μM, and with or without addition of hemin 10 μM as sole iron source. Cultures were shaken at 22°C, and growth (OD600) was measured after 12 and 24 h.

Nucleotide sequence accession number

The EMBL accession number for the sequence described in this article is AM712659.

Results and discussion

Identification of a 9-gene cluster in A. salmonicida containing putative heme uptake genes

The FURTA (Stojiljkovic et al. 1994) was used to isolate a collection of Fur box-containing clones from a genomic library of the A. salmonicida subsp. salmonicida ACR168.1 chromosome. We selected a clone which was named pFMON12 (Fig. 1a), and it contained an insert encoding a partial protein (HutA) which showed 100% identity to the A. salmonicida putative heme receptor previously sequenced (Ebanks et al. 2004) (Table 2). A second FURTA-positive clone, termed pFMON42 (Fig. 1a) contained two complete open reading frames (ORFs) encoding two putative heme utilization proteins (HutX and HutZ), and a partial ORF of a periplasmic ABC-hemin transporter protein (HutB) (Table 2). A cosmid library of A. salmonicida 168.1 previously constructed (Najimi et al. 2008) was subjected to colony-PCR-screening using PCR primers targeted to the inserts contained in pFMON12 and pFMON42. Interestingly, a single cosmid clone termed pGMON5 proved to contain the sequences of both pFMON12 and pFMON42, indicating that those genes are linked in the A. salmonicida chromosome. A DNA region spanning ca. 8,600 bp was sequenced and nine linked ORFs were identified (Fig. 1a).
Table 2

Proteins with homology to products of the A. salmonicida subsp. salmonicida ACR168.1 heme uptake cluster

A. salmonicida ORF (no. aa, KDa)

Accession no.

Identity (%)

Similarity (%)

HutA (697, 77)

Putative heme receptor (A. salmonicida)

AAS13462

100

100

Putative heme receptor (A. hydrophila)

ABK36663

92

96

Heme receptor HuvA (Listonella anguillarum)

CAC28362

29

48

ORF1 (327, 35.7)

Hypothetical iron-regulated protein (A. hydrophila)

ABK36280

83

87

Hypothetical iron-regulated protein (Vibrio angustum)

ZP_01235156

47

65

Hypothetical iron-regulated protein (Photobacterium profundum)

EAS42622

49

63

ORF2 (183, 20.1)

Isochorismatase family protein (A. hydrophila)

ABK37801

86

90

Isochorismatase family protein (Marinomonas sp.)

EAQ67201

49

70

Isochorismatase family protein (Erwinia carotovora)

CAG74672

40

63

ORF3 (142, 15.8)

Putative lipoprotein (A. hydrophila)

ABK38437

88

95

Putative lipoprotein (Pseudoalteromonas atlantica)

YP_663780

32

47

Putative lipoprotein (Shewanella baltica)

ZP_01782276

30

43

HutZ (185, 20.8)

HutZ (A. hydrophila)

ABK37918

98

98

Heme binding protein HutZ (V. cholerae)

EAX59380

67

81

HuvZ protein (L. anguillarum)

CAD43039

46

81

HutX (170, 19)

HuvX protein (A. hydrophila)

ABK36469

96

99

HuvX protein (L. anguillarum)

CAD43040

58

76

HugX (Plesiomonas shigelloides)

AAG23393

49

66

HutB (279, 29)

Periplasmic hemin transporter (A. hydrophila)

ABK36026

93

95

HuvB protein (L.anguillarum)

CAF25487

47

67

HutB (V. cholerae)

AAB94547

50

66

HutC (342, 35)

ABC transporter permease (A. hydrophila)

ABK39547

95

96

HuvC protein (L. anguillarum)

CAF25488

63

77

Hemin transporter, permease (V. parahaemolyticus)

BAC61765

61

74

HutD (262, 28.7)

ATP-binding hemin transporter (A. hydrophila)

ABK39024

90

94

ABC transporter, ATP binding protein HutD (V. cholerae)

EAY35380

50

67

ABC-type hemin transporter system, ATPase component (V. vulnificus)

NP_763478

48

65

Predicted protein sequences

The deduced amino acid sequences of six of the nine proteins encoded within this cluster (encoded by genes hutAZXBCD) shared significant degrees of similarity with known or predicted heme transport and heme utilization proteins in other bacteria (Table 2). The first gene of the cluster, hutA, encodes the predicted heme receptor protein of A. salmonicida. HutA is 100% identical to an A. salmonicida putative heme receptor previously sequenced in a proteome analysis (Ebanks et al. 2004), and shows similarity to a series of well-described outer membrane heme receptors.

A gene which we termed orf1 was found upstream of hutA and is transcribed from the opposite strand. It encodes a protein homologous to hypothetical iron-regulated proteins, but the involvement of this family of proteins in iron uptake remains unknown. orf2 encodes a putative isochorismatase-family protein. Isochorismatase genes are usually found clustered with genes that encode the biosynthetic machinery of catechol-type siderophores (Crosa and Walsh 2002), but are rarely associated with heme utilization genes. ORF3 showed similarity to putative lipoproteins. It is interesting to note that Reidl and Mekalanos (1996) demonstrated that a lipoprotein, e(P4), is essential for the utilization of hemin as porphyrin source in Haemophilus influenzae. However, ORF3 has no significant similarity with e(P4), and the relationship of ORF3 homologues with heme uptake is unknown.

HutZ is homologous to putative heme utilization proteins: it was suggested that Plesiomonas shigelloides HugZ could be involved in preventing heme toxicity (Henderson et al. 2001). In V. anguillarum, HutZ is essential for heme iron utilization (Mouriño et al. 2004). Similarly, V. cholerae HutZ is required for heme utilization as an iron source, and it is suggested that this protein acts as a heme storage protein as it binds heme with high efficiency (Wyckoff et al. 2004). The sixth ORF in the cluster corresponds to hutX. None of the HutX homologues has a well-known function, and although it has been suggested that Plesiomonas shigelloides HugX could be involved in preventing heme toxicity (Henderson et al. 2001), deletion of huvX in V. anguillarum showed to have no significant effect on the utilization of heme as an iron source (Mouriño et al. 2004).

The three remaining genes hutBCD encode proteins which show characteristics of heme transporters (Table 2). HutB is homologous to periplasmic heme binding proteins that transport heme across the periplasm from the receptor to the ABC transporter located in the inner membrane. HutC has homology to ABC-type permease proteins, and HutD shows homology to the ATP-binding protein component of ABC transporters involved in heme transport. HutD contains features common to ABC transporter ATPases (Walker et al. 1982), as the walkerA nucleotide binding consensus motif GPNGAGKS and an ABC transporter signature motif, LSGGE (data not shown), and presence of these features suggests that A. salmonicida HutD is the ATPase component of the heme ABC transporter.

Transcriptional organization of the heme uptake cluster

Since the inserts cloned into pFMON12 and pFMON42 yielded a positive result in the Fur titration assay, we would expect the existence of Fur binding sites (Fur boxes) upstream of the hutA gene and also in the intergenic region between hutX and hutB. As expected, 74 positions upstream of the putative ATG start codon of hutA we found a sequence (GAaAATGAgAATgATTccC) which shared 14 of 19 nucleotides with the E. coli consensus Fur box (de Lorenzo et al. 1987) (positions that match the consensus are shown in capitals). 69 positions upstream of hutB ATG start codon, another sequence (GATAATcgcttTCATTcTC) matched 13 positions out of 19 with the E. coli consensus (see Fig. 1a for location of predicted Fur boxes). The intergenic region between orf3 and hutZ shows a palindromic sequence (GCAATAAAAAACGGGCAGCCTGGGCTGCCCGTTTTTTATTGC) that constitutes a potential transcriptional terminator (Fig. 1a).

As shown in Fig. 1a, there are four genes or groups of genes which are transcribed from different orientations. This organization would suggest that there exist a minimum of four transcriptional units. To confirm this hypothesis, the cotranscription of orf123 was assayed in a reverse-transcription reaction with primer RT-1 targeted to the 3′-end of orf3, and PCR reactions were performed with primers that amplify orf1, orf2 and orf3 respectively. RT-PCR worked for orf2 (Fig. 1b), but no RT-PCR product could be obtained for orf1 when RT was carried out with primer RT-1. However, when we used primer RT-2 for reverse-transcription, PCR bands of the expected size were obtained for orf1 (Fig. 1b). This suggests that orf1 is independently transcribed, whereas orf2-orf3 are cotranscribed.

Similarly, to assay the cotranscription of hutX-hutZ on the one side, and of hutBCD on the other side, reverse-transcription was carried out with primers RT-3 and RT-4 respectively, and subsequent PCR reactions were conducted with primer pairs targeted to hutX and hutB respectively. Each RT-PCR yielded a product of the expected size (Fig. 1b), and negative and positive controls corroborated the accuracy of the RT-PCR experiment. These results altogether indicate that the nine genes of the described cluster are organized into five transcriptional units, three polycistronic ones: orf2-orf3, hutX-hutZ, hutBCD and two monocistronic ones, hutA and orf1, which would be each transcribed from their own promoters.

Similarities to other heme transport gene clusters

The spatial organization of heme uptake cluster in the chromosome of A. salmonicida showed some similarity to other described heme uptake gene clusters in Gram-negative bacteria, although some differences exist (Fig. 2). The gene encoding the outer membrane heme receptor is linked to the rest of the heme transport genes in A. salmonicida and A. hydrophila, a characteristic shared with V. anguillarum (Mouriño et al. 2004). However, in the majority of the Vibrio species (we include V. cholerae as an example in Fig. 2) the outer membrane receptor is not linked to the heme cluster but rather located in another chromosomal location. In addition, some species as V. cholerae, Photobacterium damselae and Plesiomonas shigelloides, contain within their heme uptake gene clusters a gene which is transcribed from the same strand as hutXZ (Occhino et al. 1998; Rio et al. 2005; Henderson et al. 2001) but this gene is absent in the A. salmonicida cluster. This gene codes for a putative coproporphyrinogen oxidase (HutW in V. cholerae) an enzyme that converts coproporphyrinogen III into protoporphyrin IX (Panek and O’Brian 2002). In A. salmonicida as well as A. hydrophila, three genes (orf1, orf2, orf3) which encode a hypothetical iron regulated-protein, an isochorismatase-family protein and a putative lipoprotein are linked to the heme cluster genes (Seshadri et al. 2006), but homologues of these genes have not been reported in the well-described heme uptake clusters of related Gram-negative bacteria. It is also interesting to note that many Gram-negative bacteria contain genes of a TonB system linked to the heme transport genes, whereas no TonB system genes are linked to the heme uptake genes in A. salmonicida (Fig. 2). The fact that these two last characteristics are shared uniquely with A. hydrophila suggests that this type of gene organization in the heme uptake cluster could be a feature of the genus Aeromonas.
https://static-content.springer.com/image/art%3A10.1007%2Fs00203-008-0391-5/MediaObjects/203_2008_391_Fig2_HTML.gif
Fig. 2

Comparative arrangement of heme uptake cluster genes of A. salmonicida, A. hydrophila, V. anguillarum and V. cholerae. Note that in the two Aeromonas species there are no homologous of tonBexbBD genes linked to the rest of heme transport genes

The A. salmonicida putative heme transport genes allow E. coli 101ESD to utilize heme and hemoglobin as sole iron sources

In order to demonstrate that the genes of this cluster play a role in heme transport, we utilized as a heterologous host the E. coli strain 101ESD. This strain lacks a genetic system for heme uptake and does not produce siderophores, and thus it cannot grow in the presence of iron chelators unless supplied with an utilizable source of iron. Genes encoding heme transport functions can thus be cloned into plasmids and tested whether they provide E. coli 101ESD with the ability to grow under conditions where heme and/or hemoglobin are the only iron sources available.

Results of the experiments carried out with E. coli 101ESD are summarized in Table 3. As a positive control, we used E. coli 101ESD transformed with cosmid pML1 (Table 1) that contains the whole heme uptake system of V. anguillarum (Mazoy et al. 2003; Mouriño et al. 2004). This transformant is able to utilize hemin and hemoglobin as iron sources. A. salmonicida hutA alone (pMON34) could not complement E. coli 101ESD, indicating that other genes in addition to hutA are necessary for the utilization of heme compounds as iron sources. We then cloned the whole A. salmonicida 9-gene cluster into vector pACYC184 to yield pMON33 and transformed it into E. coli 101ESD. Surprisingly, this transformant was unable to grow in the presence of hemin and hemoglobin as iron sources. This would suggest that, either additional genes are necessary, or that some of the genes provided in pMON33 are not functional. We thus attempted to demonstrate whether hutA actually functions as an outer membrane heme transporter. For this purpose, we transformed E. coli 101ESD with plasmid pSML33 (Table 1) that contains all the heme transport genes of V. anguillarum with the exception of huvA (encoding the V. anguillarum heme receptor) and does not confer the ability to use heme sources unless combined with a plasmid encoding a functional outer membrane heme transporter (Mouriño et al. 2004, 2005). We thus introduced plasmid pSML33 + plasmid pMON34 in 101ESD. This strain was able to grow using both hemin and hemoglobin, demonstrating that A. salmonicidahutA is indeed a functional outer membrane heme transporter (Table 3).
Table 3

Utilization of hemin and hemoglobin as iron sources by E. coli 101ESD complemented with different combinations of A. salmonicida heme uptake genes, as well as heme uptake genes from Vibrio anguillarum and Photobacterium damselae

Plasmid combination

Gene(s) present

Hemoglobin (1 mM)

Hemin (5 mM)

FeSO4 (5 mM)

101ESD

None

a

+a

101ESD/pML1

huvAZXtonBexbBDhuvBCD

+

+

+

101ESD/pMON33

hutA-orf1-orf2-orf3-hutZX-hutBCD

+

101ESD/pMON33/pCAR168

hutA-orf1-orf2-orf3-hutZX-hutBCD + tonBexbBD

+

+

+

101ESD/pCAR168

tonBexbBD

+

101ESD/pSML33

tonBexbBDhuvBCD

+

101ESD/pMON34

hutA

+

101ESD/pMON34/pSML33

hutA + huvZX tonBexbBD-huvBCD

+

+

+

101ESD/pMON36/pCAR168

hutAZX + tonB-exbBD

+

101ESD/pMON37

hutBCD

+

aGrowth of E. coli 101ESD around the disks supplemented with hemoglobin, hemin or FeSO4 was annotated as positive (+) or negative (−) after 24 h of incubation

We hypothesized that one of the reasons why pMON33 did not complement 101ESD could be the inability of the E. coli TonB system to function with the A. salmonicida heme receptor. To test this possibility, we transformed 101ESD with both pMON33 and with pCAR168, a plasmid that contains the tonBexbBexbD genes from the fish pathogenic bacterium Photobacterium damselae (Table 1). Interestingly, this new gene combination allowed 101ESD to utilize both hemin and hemoglobin as iron sources, indicating that this nine-gene cluster cloned into pMON33 is functional for the transport of heme groups as long as a functional TonB system is provided.

Although no tonbexbBD genes are present within the cluster described in this study, we would expect that homologous of these genes might be located at another chromosomal location in A. salmonicida. While this study was being completed, the genome sequence of A. salmonicida A449 was made available in public databases (GenBank Accession number CP000644). An in silico search in this genomic sequence revealed the existence of genes for two different TonB systems (TonB1 and TonB2) which are not linked to the heme uptake cluster genes in that strain (data not shown).

When the A. salmonicida heme uptake cluster with the exception of hutBCD genes (pMON36, Fig. 1b) was transformed into 101ESD containing pCAR168, the transformant failed to utilize heme iron sources. This indicates that an outer membrane receptor and a TonB system genes are not sufficient for heme iron utilization, the genes encoding the predicted periplasmic and ABC-transporter genes of A. salmonicida being also necessary. Complementation of E. coli strains with heme uptake genes provided in trans to allow heme uptake has been previously reported with several genetic systems (Stojiljkovic and Hantke 1992; Mouriño et al. 2004, 2005; Rio et al. 2005). Similarly, Occhino et al. (1998) reported that utilization of hemin as an iron source can be reconstituted in E. coli 1017 (HB101 derivative, ent) with V. cholerae hutA, tonB exbBD and hutBCD genes.

We found that the 101ESD TonB system was not functional with the A. salmonicida heme transport genes. Similarly, it was reported that the Plesiomonas shigelloides heme receptor gene hugA could not complement an E. coli DHE-1 hemA mutant for the use of heme as a porphyrin source unless Plesiomonas shigelloides tonB-exbB-exbD genes were also provided in trans (Henderson et al. 2001).

Altogether, the results obtained in this section demonstrate that hutA is a functional outer membrane heme transporter, that the gene operon constituted by hutBCD is also necessary for heme uptake in E. coli 101ESD, and that this nine-gene cluster of A. salmonicida contains all the genetic determinants with the exception of a TonB system, as to allow E. coli 101ESD to use heme compounds as iron sources.

Deletion of hutA and hutB genes affects the ability of A. salmonicida to grow with hemin as sole iron source

We have demonstrated that this nine-gene cluster contains determinants for the utilization of heme groups as iron sources in E. coli 101ESD. We thus aimed at elucidating the role of these transport genes in A. salmonicida. For this purpose, in-frame deletions of hutA and hutB were constructed. We selected MON15, a RSP74.1 derivative strain, rather than ACR168.1 because this last strain is unable to grow in minimal medium, likely due to an uncharacterized auxotrophy. MON15 also harbors the heme uptake cluster described above, sharing more than 99% identity in nucleotide sequence with sequence of ACR168.1 (data not shown). In addition, MON15 is an ΔasbD mutant deficient in siderophore biosynthesis (Najimi et al. 2008), and it constitutes an ideal model to study the utilization of heme as a unique iron source because the background growth resulting from siderophore-mediated iron uptake is eliminated.

No significant differences in growth rates were seen between the hutB mutant (MON30) and A. salmonicida MON15 when grown in CM9 for 12 or 24 h (Fig. 3). The same strains were then grown in an iron restricted medium with 2,2’-dipyridyl added at a concentration of 60 μM, and both the parental and mutant strains were significantly impaired for growth (Fig. 3). When hemin was added at a concentration of 10 μM in addition to 2,2’-dipyridyl, the parental strain MON15 significantly recovered the ability to grow, showing that this strain is able to utilize heme as the sole iron source. However, under these same conditions, the hutB mutant was affected in its ability to use hemin, showing more than a fivefold reduction in growth levels after 24 h, compared to the parental strain (Fig. 3b). The phenotype of the putative periplasmic heme transporter hutB defective mutant suggests that this gene is necessary, if not totally essential, for growth with hemin as sole iron source in the conditions described in this study. The residual growth of the hutB mutant in CM9 + hemin, could be attributed to other periplasmic iron-transporters that may be encoded in the A. salmonicida genome. In other species, individual genes encoding periplasmic hemin-binding proteins and hemin ABC-transporters demonstrated not to be essential for heme uptake. As an example, mutation of either the hutB or the hutC homologues in V. cholerae only cause a modest decrease in the ability of the bacterium to use heme as iron source. This observation led the authors to suggests that additional ABC transporter genes could play a redundant role in the uptake of hemin (Occhino et al. 1998).
https://static-content.springer.com/image/art%3A10.1007%2Fs00203-008-0391-5/MediaObjects/203_2008_391_Fig3_HTML.gif
Fig. 3

Growth (OD600) after 12 h (a) and 24 h (b) of incubation, of A. salmonicida MON15 (parental strain), and the mutant strains MON29 (ΔhutA) and MON30 (ΔhutB), under the following three different conditions: CM9 (iron sufficient conditions), CM9 + 60 μM EDDA (iron-limiting conditions) and CM9 + 60 μM EDDA + 10 μM hemin (to test use of hemin as sole iron source). Results are mean values from three independent experiments and error bars are shown

We then analyzed the effect of the hutA mutation in A. salmonicida. When the ΔhutA mutant (MON29) was grown in CM9, or in CM9 + 60 μM 2,2’-dipyridyl, it did not show significant differences with the parental strain after 12 or 24 h (Fig. 3a, b). Interestingly, when the ΔhutA strain was grown with 10 μM hemin it showed some defect in its ability to grow with this unique iron source during the first 12 h (Fig. 3a), but recovered growth levels similar to the parental strain after 24 h of culture (Fig. 3b). This result indicates that hutA is not totally essential for growth of A. salmonicida with hemin as the sole iron source, and suggests that additional outer membrane heme receptor genes exist in the genome of A. salmonicida. Presence of more than one functional outer membrane heme receptor gene coexisting in the same genome has been reported for other gram negative bacteria. This is the case of V. cholerae, where at least three functional heme receptors, namely HutA, HutR and HasR contribute to heme iron acquisition in this pathogen (Mey and Payne 2001), and mutation of any of these three receptors do not abolish completely the ability of the bacterium to use heme as iron source, being necessary the inactivation of the three genes in order to abolish hemin utilization. In V. anguillarum two different heme receptors, HuvA and HuvR have been described, although the genes for these two receptors have never been found coexisting in the same V. anguillarum strain (Mouriño et al. 2005).

An in silico search in the genome sequence of A. salmonicida A449 showed that one of the ORFs (locus ASA_2695) is annotated as an hemin receptor. Interestingly, this ORF encodes a predicted protein that shares 30% identity with the sequence of HutA described in this study, and is 37% identical to V. cholerae HutR (Mey and Payne 2001). This gene might thus encode an alternate outer membrane heme receptor. Mutational analysis of this putative heme receptor gene together with the construction of double mutants in combination with hutA, will allow a deeper insight into the genetic basis of heme transport in this important fish pathogen.

Acknowledgments

This work was supported by Grant PGIDIT06RMA26101PR-2 and Contract No. 2004/CP481, from Xunta de Galicia, Spain. Mohsen Najimi acknowledges the Ministry of Science and Education of Iran for a predoctoral fellowship.

Copyright information

© Springer-Verlag 2008