Evidence of Siderophore and Iron-Rep

Advances in Biological Research 1 (3-4): 122-129, 2007 ISSN 1992-0067 © IDOSI Publications, 2007

The Response of a Siderophore-Degrading Bacterium (Mesorhizobium loti) to Iron-Deprivation: Evidence of Siderophore and Iron-Repressible Protein Synthesis Jodi Morton, Katherine Marsh, Megan Frawley and Domenic Castignetti Department of Biology, Loyola University of Chicago, 6525 North Sheridan Road, Chicago, IL 60626 Abstract: Previous study demonstrated that a Mesorhizobium loti strain can use the siderophore deferrioxamine B as its sole source of carbon for growth. The cognate ferrisiderophore (ferrioxamine B), however, is unable to supply the M. loti with iron. The bacterium must thus invoke an alternate mechanism to gain the iron it requires for growth. Data is presented which indicates that the M. loti synthesizes its own siderophore, trivially named lotibactin, which displays no distinct ferrated, versus non-ferrated, ultra violet-visible spectra but which competes successfully with a known ferric ion chelator, citric acid, for ferric ion. Typical of many Gram negative bacteria, the M. loti also responds to iron-deprivation by invoking iron-repressible outer membrane proteins. Key words: Siderophore

Mesorhizobium loti

iron assimilation

INTRODUCTION

Determining the presence of the latter type of siderophore has been done primarily via the Chrome Azurol S (CAS) assay of Schwyn and Neilands [16]. This assay results in an unknown structure of a Fe3+-CAShexadecyltrimethylammonium (HDTMA) complex, at pH 5.6, that changes from a blue to an orange-yellow color as the Fe3+ is removed from the complex [2, 16]. Data concerning the stability (formation) constants of the Fe3+-CAS chelate were collected at pH 2-4 [17] without the detergent HDTMA being present. Langmyhr and Klausen [17] determined that five different Fe3+-CAS complexes were formed, that is, the 2-, 1 -, the neutral, the 2 + and a complex of unknown charge. The stability constants of these complexes ranged from 36.2 for the 2-complex, 15.6 for the 1- and neutral complexes and 20.2 for the 2+ complex. A search of the literature yielded no known stability constant of the Fe3+-CAS-HDTMA complex. In addition to the uncertain nature of the strength of Fe3+ binding by the CAS-HDTMA complex, the CAS is also subject to interferences. For example, Winkelmann et al. [18] reported that growth medium components interfered with the CAS assay, rendering it unsuitable for determining the presence of a siderophore in the culture medium of the bacterium they were examining and that phosphate, a common buffer component, also interferes with the assay [11]. Our laboratory has been investigating the metabolism of a soil bacterium, a Mesorhizobium loti, that has the

Siderophores are microbially-synthesized, highaffinity ferric ion (Fe3+) chelators [1-5]. Due to the near insolubility of iron in environments at, or near, neutral pH and where oxygen is present, microbes elicit siderophores to chelate the sparse quantities of Fe3+ present. The Fe3+-laden siderophore (a ferrisiderophore), then delivers the Fe3+ it contains to the microbial cell, where it is assimilated via an energy-dependent, highly specific process [4, 6-8]. In Gram negative bacteria, this response results in specific outer membrane proteins [Iron-repressible Outer Membrane Proteins (IROMPs)] being synthesized whose role is to recognize and assimilate the ferrisiderophore through the organism’s outer membrane. Siderophores often belong to either the hydroxamate or catecholate molecular classes [1, 3, 7]. Such siderophores may be ascertained by directly noting the ability of these compounds to react in chemical assays specific for either the catecholate or hydroxamate moieties and by the presence of characteristic ferrisiderophore or deferrisiderophore ultraviolet (UV) or visible (VIS) spectra [9-11]. Siderophores which lack these chemical moieties, however, have been noted [12-15]; indeed the first such identified siderophore [13, 14] not only lacked either catecholate or hydroxamate ligands but also had no distinguishing UV or visible absorbencies.

Corresponding Author: Dr. D. Castignetti, Department of Biology, Loyola University of Chicago, 6525 North Sheridan Road, Chicago, IL 60626

122

Advan. Biol. Res., 1 (3-4): 122-129, 2007

ability to catabolize a siderophore (deferrioxamine B-DFB) and use it as a source of carbon for growth [19-22]. Two other such microbes are known [23-25] to possess the ability to catabolize siderophores [18, 26] but to date, whether these microbes synthesize their own specific siderophores has not been reported. When the M. loti is grown in the medium which contains DFB as its sole source of carbon, the siderophore is far in excess of its cognate ferrisiderophore, ferrioxamine B (FB). We began the investigation of how M. loti acquires the Fe3+ it requires for growth by noting whether it could use FB as an iron source. Examination of M. loti’s ability to assimilate the Fe3+ of FB revealed that it was incapable of doing so, although other Fe3+-chelators were able to supply the metal [27]. The M. loti most likely produces its own siderophore in order to compete with the FB present in the DFB-FB containing growth medium. Preliminary experiments revealed that growth of the M. loti in a synthetic medium, in which Fe3+ has been removed, resulted in the synthesis of a molecule (s) that was CAS positive but whose UV-visible absorbance, whether ferrated or not, was not distinctive. This result led us to further characterize M. loti’s response to Fe3+-deprivation by developing a facile Fe3+-citrate (Fecitrate) based spectral assay. As ferric-citrate (as either the mono-Fe3+-monocitrate or di-Fe3+-dicitrate complex) has characteristic formation constants (11.5 and 22.1, respectively-Sillen and Martel 1964; 1971), the chelation capacity of the M. loti siderophore, which we have termed “lotibactin,” can be correlated to a known standard. In addition, we attempted to correlate the synthesis of lotibactin with IROMPs in the M. loti outer membrane. MATERIALS AND METHODS As extraneous siderophores would complicate lotibactin characterization, it was prudent to have no other siderophores (such as its carbon source, DFB) present in the growth medium. We thus developed a minimal medium in which the concentration of iron was kept low due to its removal by the resin Chelex 100 (BioRad, Richmond, CA). The medium consisted of a mineral salts medium, minus DFB and FeCl3, which was previously described [19]. It contained a 60 mM phosphate buffer at pH 7.0 and trace minerals, each at 0.5 mg l 1, of CaSO4-2H2O, CuSO4-5H2O and ZnSO4-7H2O. The mineral base was sterilized via autoclaving and to it were added, per liter, 0.5 g NH4Cl, 0.5 g MgSO4-7 H2O, 2 g fructose, 2 g Na-acetate and 0.5 ml 123

of 10X vitamins [28], all of which had been sterilized by filtration (0.22 µm filter) prior to being added to the mineral base. These components had been deferrated by passing them through a Chelex 100 (~40-50 ml) column that was used in its Na-form. After sufficient cellular growth of M. loti cells, either wild-type or hyper-producing siderophore mutants, the supernatants were collected via centrifugation (10,000 x g for 10 min) and were verified as being CAS assay [16] positive before subsequent steps were taken. The siderophore hyper-producing mutants were obtained via the manganese selection technique [29] which selects cells deficient in the regulation of siderophore synthesis and hence overproduce siderophores. All chemicals, excluding the chromatography resins, were purchased from Fisher Scientific Company (Hanover Park, IL) or Aldrich Chemical Company (Milwaukee, WI) and were of ACS, or higher, quality. Partial purification of lotibactin was achieved by loading CAS-positive supernatant (2 to 40 ml of resin) onto a Sephadex A-50 (Pharmacia, Piscataway, NJ) column, the resin being suspended in a 10 mM phosphate buffer, pH 7.0. Lotibactin was eluted by adding a 0.5 M NaCl-10 mM phosphate buffer, pH 7.0 and 1 ml fractions were collected and tested for CAS positive reactions. These fractions were then pooled together. To initiate the ferric citrate spectral assay, a 2 mM Ferric-citrate-10mM PO4, pH 7.0, stock solution was made. Differing volumes of either the lotibactin-containing culture supernatant, or the Sephadex A-50-CAS positive eluate, were brought to a total volume of 500 µl with 10 mM phosphate buffer, pH 7.0. These volumes were then mixed, 1:1, with 500 µl of the 2 mM ferric-citrate-10 mM phosphate solution, yielding a 1 mM ferric-citratelotibactin solution. An iron-laden lotibactin (ferrilotibactin) was also prepared to use as a control by mixing 20 µl of 100 mM FeCl 3-10 mM HCl with either 500 µl of lotibactin from culture medium or from the Sephadex A-50 eluate and 1.48 ml of the 10 mM phosphate buffer. In the assay, EDTA was employed as the standard for a strong ferric ion-chelator (stability constant of 25.1) [30,31]. Alanine and glucose were employed as the standards for weak ferric ion-chelators (stability constants of 11.0 and not reported, respectively) [30, 31]. EDTA, alanine and glucose were used in a 0-20mM concentrations and mixed with 2mM Fe-Citrate to yield solutions of 0-10mM EDTA, glucose or alanine with 1mM Fe-Citrate.

Advan. Biol. Res., 1 (3-4): 122-129, 2007

that M. loti cells grown in the presence of iron be collected by centrifugation (10,000 x g for 10 min) and washed 5 times in 10 mM phosphate buffer, pH 5.6. This step was necessary to help dissolve and remove the metal from the cells in order that samples suitable for analysis could be obtained. M. loti was grown in 200 ml of the medium listed above to promote the synthesis of its siderophore (lotibactin) and putative IROMPs. Cells were harvested at mid-late log phase by centrifugation at 15,000 x g for 15 min at 4 C. At this point, M. loti cells grown in ironreplete conditions (1 mM) were washed five times in 10 mM phosphate buffer, pH 5.6. Pellets, from either iron-deficient or replete medium, were then resuspended in 8 ml of 0.75 M sucrose buffer described by Stull et al. [33] with the exception that 800 µg nuclease P 1 from Penicillium citrinum (Calbiochem-Behring DiagnosticsLaJolla, CA) was substituted for the 800 µg of DNase and 800 µg of RNase described in the Stull et al. [33] procedure. Cells were sonicated 6 times, each time for 30 sec with sufficient time (~30 sec) between sonications for the cells to chill before the next sonication was begun. The solution was next centrifuged at 1500 x g for 15 min, the pellet was removed and the supernatant was stored overnight at 4°C. Centrifugation at 186,000 x g for 1 hour collected the crude membranes and this pellet was resuspended in 1 ml of 10 mM Tris-acetate, pH 7.8, buffer and an equal volume of 4% Trition X-100. This solution was left at room temperature (~24°C) for 20 min and then centrifuged at 186,000 x g for 1 hour. The resulting pellet was resuspended in 2 ml of 2% Trition X-100-10 mM Tris-acetate, pH 7.8 and incubated for an additional 20 min at room temperature. The mixture was centrifuged again at 186,000 x g for 1 hour and the pellet was resuspended in a small volume of distilled, de-ionized water (0.2-1.5 ml, depending on pellet size) and stored at-70 C until used. Prior to adding 2X Lugtenberg buffer, as listed by Hoefer Scientific Instruments (Hoefer SE 400 Sturdier Slab Gel Electrophoresis Unit Manual-San Francisco, CA), the protein content of the pellet was measured using the bichinchonic acid procedure of Pierce (Rockford, IL) and the protein content was adjusted to 2 mg protein per ml before mixing with an equal volume of the 2X Lugtenberg buffer. Samples were then loaded onto the apparatus as described by its manufacturer (Hoefer Scientific Instruments). SDS-PAGE gel electrophoresis was performed on the Hoefer SE 400 Sturdier Slab Gel Electrophoresis Unit (Hoefer, San Francisco, CA) using either 9 or 10% gels, prepared and used as described by the manufacturer.

Ferrated solutions of EDTA, alanine and glucose were used as controls. These solutions were made by adding 20 µl of 100 mM FeCl3-10 mM HCl, to 1.48 ml of 10 mM phosphate buffer and to 500 µl of the 0-20 mM EDTA, alanine, or glucose (in 10 mM phosphate buffer) solutions. These controls were performed to note any spectra of either ferri-EDTA, ferri-glucose, ferri-alanine, or ferrilotibactin as compared to a Fecitrate spectrum. None of these solutions interfered with the Fecitrate absorption spectrum. Other control solutions were 1 mM Fecitrate, in 10 mM phosphate-500 mM NaCl (as would be present in Sephadex A-50 eluted and mixed samples), demonstrated an identical absorption spectrum as did 1 mM Fecitrate in 10 mM phosphate. An additional control of citric acid, without ferric ion being present (1 mM citric acid in 10 mM phosphate buffer, pH 7.0), was made and noted not to interfere with the assay. Controls for the presence of the M. loti growth medium components (that is, primarily the elevated concentration of phosphate along with fructose, sodium acetate and the mineral components present in the growth medium), yielded spectra that were nearly identical to that of the 1 mM Fecitrate-10 mM phosphate buffer spectrum, with only a minor shift toward minimally larger absorbencies in the 300-350 nm range observed. All solutions were allowed to equilibrate for at least 12 hours at 4°C prior to being brought to room temperature for spectral analysis. On those occasions where a precipitate was formed, it was removed via centrifugation prior to performing the spectral analysis. Spectra were recorded on a Cary 100 spectrophotometer (Varian, Walnut Creek, CA), measuring absorbencies between 190-500 nm. The baseline was established using a 10 mM phosphate buffer, pH 7.0. Control and the individual sample, spectra were confirmed as accurate by using freshly made samples and performing the analyses on at least two different occasions. To assess whether the M. loti responded to ironstarvation by coordinating siderophore production with the insertion of IROMPs into its outer membrane, the M. loti was grown to mid-to late-logarithmic stage in the medium described above, either with Fe3+ (final concentration of 1 mM FeCl 3-6H2O) or without Fe3+ being added. The production of siderophore in these cultures was either confirmed (iron-starved conditions) or noted as absent (iron-replete conditions) via use of the CAS assay [16]. The procedures employed to isolate and identify M. loti’s IROMPS included those that specifically isolated the outer membrane [8, 32, 33] as well as those that collect the membranes from the whole cell [34-37]. The procedure which proved successful, that of Stull et al. [33], required 124

Advan. Biol. Res., 1 (3-4): 122-129, 2007

Standards used, each at 1 mg ml 1 in the Tris-SDSglycerol-2-mercaptoethanol-0.05% bromphenol blue treatment buffer described by the manufacturer, were egg albumin (45 kDa), bovine serum albumin (66 kDa), conalbumin (80 kDa) and b-galactosidase (116 kDa) and were from either Sigma (St. Louis, MO) or Fisher Scientific Company (Hanover Park, IL). Volumes of the standards used were between 4-8 µl each while the IROMP and Fe-sufficient outer membrane samples were loaded at 20-100 µl each. Data of the Rf values of the standard proteins versus their molecular weights (Mr) were obtained via SDS-PAGE gel analyses. Using Microsoft Excel, the standard protein data were used to plot a Mr versus Rf graph and a line of best fit (logarithmic) was obtained from the program. The line of best fit equation was used to generate Mr values of the protein bands present in the outer membranes of both iron-deprived and iron-replete M. loti cells. Typical correlation coefficients for the lines of best fit generated by these gels were R2 = 0.9995. RESULTS Fecitrate did not alter its ultraviolet and visible spectrum when the weaker ferric ion chelators, glucose or alanine, were the competing ligands Fig. 1. Data collected with Fecitrate and alanine are presented but are representative of the spectra obtained when either

glucose or alanine were mixed with the ferric citrate solution. In contrast, a noted spectral shift was observed when the superior chelator, EDTA was present Fig. 2. As EDTA is a superior ferric ion chelator compared to citric acid [30,31], the alteration of the Fecitrate spectrum by ferric-EDTA is hypothesized to mimic what a siderophore with either no, weak or a poorly defined visible-ultraviolet spectrum would do to the Fecitrate spectrum. In contrast, the siderophore deferrioxamine B, a strong ferric ion chelator with a well-defined spectrum, readily removed the ferric ion of the Fecitrate complex and the characteristic ferrioxamine B spectrum was formed (data not shown). When the putative siderophore lotibactin was mixed with Fecitrate, it altered the Fecitrate spectrum in a manner analogous to that observed with EDTA Fig. 3. A point of note is that lotibactin did not display markedly different and distinguishable ultra violet-visible spectra when ferrated as compared to when unferrated. This is in contrast to what is typically noted of either hydroxamate or catecholate siderophores but is similar to what has been observed with complexone-type siderophores as the latter molecules lack charge transfer bands [12-15]. To correlate the production of M. loti’s putative siderophore with IROMPs normally associated with iron starvation of Gram negative bacteria, we attempted to isolate and determine the Mr of the outer membranes present during iron starvation but absent, or markedly reduced, in quantity when iron was replete.

4 3 2 1 0

200

300

Wavelength (nm)

400

500

Fig. 1: Visible and ultra violet absorption spectra of 1 mM Fecitrate and 1 mM Fecitrate in the presence of 1, 2, 5 and 10 mM alanine. The 1 mM Fecitrate (with no alanine present) spectrum and the spectra of the 1 mM Fecitrate samples with 1, 2, 5 and 10 mM alanine are super-imposable. For purposes of clarity, the spectra of the controls (20 mM alanine in the absence of ferric ion, 1 mM citric acid, 10 mM alanine, 0 mM alanine, 4 mM alanine and 8 mM alanine) are not shown as all had absorbance maxima of less than 2 and all had essentially an absorbance of 0 at wavelengths greater than 225 nm. All samples were prepared in the 10 mM phosphate buffer (pH 7.0) and were scanned against a 10 mM phosphate buffer (pH 7.0) blank (the reference blank) 125

Advan. Biol. Res., 1 (3-4): 122-129, 2007

4 3 2 1

200

300

Wavelength (nm)

400

500

Fig. 2: Visible and ultra violet absorption spectra of 1 mM Fecitrate, 1 mM Fecitrate in the presence of 1, 2, 5 and 10 mM EDTA and 1 mM Fe-EDTA samples. The top 4 lines are ferric chloride-EDTA controls, are superimposable and consist of 1 mM FeCl3 in 10 mM phosphate buffer with 2, 4, 10 and 20 mM EDTA. The next 4 lines, also superimposable, are 1 mM Fecitrate in10 mM phosphate buffer with 1, 2, 5 and 10 mM EDTA. The next line is 1mM Fecitrate-0mM EDTA. The bottom line is 1 mM FeCl3 with 0 mM EDTA. All samples were prepared in the 10 mM phosphate buffer (pH 7.0) and were scanned against a 10 mM phosphate buffer (pH 7.0) blank (the reference blank) 5 4 3 2 1 0

200

300

Wavelength (nm)

400

500

Fig. 3: Visible and ultra violet absorption spectra of Fecitrate and Fecitrate with lotibactin. The top line at approximately 350 nm is 1 mM Fecitrate containing 500 µl of lotibactin supernatant. The second through sixth lines are, respectively, 1 mM Fecitrate containing 400 µl of lotibactin supernatant and 100 µl of 10 mM phosphate buffer, 1 mM Fecitrate containing 300 µl of lotibactin supernatant and 200 µl of 10 mM phosphate buffer, 1 mM Fecitrate containing 200 µl of lotibactin supernatant and 300 µl of 10 mM phosphate buffer, 1 mM Fecitrate containing 100 µl of lotibactin supernatant and 400 µl of 10 mM phosphate buffer and 1 mM Fecitrate (these two lines are virtually indistinguishable). The next two lines are, respectively, 500 µl of lotibactin supernatant and 500 µl of 10 mM phosphate buffer and 1 mM FeCl3 with 500 µl of lotibactin supernatant, prepared as described above for ferriclegiobactin. The bottom line is 1 mM FeCl 3 in 10 mM phosphate buffer Using the Stull et al. [33] procedure, it was observed that 17 proteins of molecular weights greater than 62,000 were present in the outer membrane of cells grown in the iron-insufficient medium while 14 were present at comparable levels in iron-replete grown cells. The 3 outer membrane proteins (IROMPs) present in the iron-deprived

Six isolation procedures for cellular or outer membranes were tested before one, that of Stull et al. [33], proved satisfactory. Growth in the iron-deficient medium resulted in siderophore synthesis, as determined by the CAS assay, while growth in the iron-replete medium demonstrated no siderophore synthesis. 126

Advan. Biol. Res., 1 (3-4): 122-129, 2007

cells, but either absent or greatly reduced in quantity in the iron-replete cells, had molecular weights of 69 kDa, 94 kDa and 99 kDa.

X-100 and a Tris-acetate buffer, proved satisfactory for obtaining outer membranes from M. loti, although growth of the bacterium in an iron-replete medium, which repressed synthesis of putative IROMPs, required extensive washing of the cells to remove the iron. In this regard, our attempts to isolate the IROMPs of M. loti suggest that caution must be used before applying techniques designed for E. coli to Gram negative bacteria such as M. loti. Using the Stull et al. procedure, M. loti was observed to synthesize 3 putative IROMPs with molecular weights of 69, 94 and 99 kDa. M. loti thus responded to iron deprivation in a manner analogous to other Gram negative bacteria, which also synthesize IROMPs in this molecular weight range [4, 7]. Which of these putative IROMPs, if any, actually bind and transport ferrilotibactin into the cell must await the purification and isolation of this ferric ion-binding molecule(s).

DISCUSSION The formation constants of ferric citrate (11.5 for the mono-ferric citrate and 22.1 for the di-ferric dicitrate complex) offer a ready comparison of a compounds’ ability to chelate ferric ion. Inferior chelators, such as glucose and alanine, failed to alter the Fecitrate spectrum even when in molar excess. In contrast, the superior ferric ion chelator, EDTA, with a formation constant of 25.1, readily interacted with the Fecitrate complex and caused a notable shift in its absorption spectrum, a shift that was apparently complete at a 1:1 molar ratio of Fecitrate to EDTA. Interestingly, the spectral data displays that the Fecitrate-EDTA spectra are intermediate between that of Fecitrate and that of ferric EDTA (with no citrate present). We thus hypothesize that the observed Fecitrate-EDTA spectra may represent a joint chelation of the ferric ion by both the EDTA and citrate. The putative ferrated and unferrated siderophores of M. loti (lotibactin) displayed spectra with much of their absorbencies occurring in the deep ultra violet. While this siderophore has yet to be thoroughly characterized as to its spectral and chemical characteristics, it clearly mimicked the interaction with Fecitrate that was observed with EDTA, causing a shift to higher wavelengths versus that observed with Fecitrate alone. As with EDTA, the higher concentrations of lotibactin have a titration effect upon the curve. Like EDTA and unlike either glucose or alanine, lotibactin was superior to Fecitrate as a ferric ion chelator. Its presence in the 10 mM phosphate buffer with Fecitrate resulted in the alteration of the Fecitrate spectrum in a manner analogous to that observed with EDTA, thereby demonstrating a superior ability to chelate ferric ion versus that of citrate. Gram negative bacteria experiencing iron starvation commonly display IROMPs in their outer membranes [4, 7]. Isolation of good quality membrane preparations from our M. loti was problematic. The reasons for this perhaps derive from the procedures being designed for outer membrane isolation from Escherichia coli, which proved to be less than ideal for the isolation of IROMPs from M. loti. Modifications, such as the addition of protease inhibitors (phenylmethylsulfonofluoride, leupeptin, EDTA and pepstatin A), did little to resolve the problem of obtaining suitable samples. The Stull et al. [33] procedure, which employs nucleases, lysozyme, Triton

ACKNOWLEDGEMENTS DFB was a gift from the Novartis Corporation (Summit, NJ). This work was supported, in part, by USDA grants NRI 2002-35101-11532 and NRI 2003-35107-13886 and a Loyola University research support grant to D.C. REFERENCES 1. 2.

3. 4. 5. 6.

7. 8.

127

Emery, T., 1982. Iron metabolism in humans and plants. Am. Sci., 70: 626-631. Fekete, F., 1993. Assays for microbial siderophores. In: Iron chelation in plants and soil microorganisms. L. Barton and B.C. Hemming, Ed.s. Academic Press. San Diego, pp: 399-417. Neilands, J., 1981. Microbial iron compounds. Annu. Rev. Biochem., 50: 715-731. Neilands, J., 1982. Microbial envelope proteins related to iron. Annu. Rev. Microbiol., 36: 285-309. Neilands, J., 1995. Siderophores-structure and function of microbial iron transport compounds. J. Biol. Chem., 270: 26723-26726. Enard, C., A. Diolez and D. Expert, 1988. Systemic virulence of Erwinia chrysanthemia 3937 requires a functional iron assimilation system. J. Bacteriol., 170: 2419-2426. Guerinot, M., 1994. Microbial iron transport. Annu. Rev. Microbiol., 48: 743-772. Gensberg, K., K. Hughes and A. Smith, 1992. Siderophore-specific induction of iron uptake in Pseudomonas aeruginosa. J. Gen. Microbiol., 138: 2381-2387.

Advan. Biol. Res., 1 (3-4): 122-129, 2007

9. 10. 11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Neilands, J., 1966. Naturally occurring non-porphyrin iron compounds. Struct. Bonding., 1: 59-108. Neilands, J., 1984. Methodology of siderophores. Struct. Bonding., 58: 1-24. Neilands, J. and K. Nakamura, 1991. Detection, determination, isolation, characterization and regulationof microbial iron chelates. In: CRC Handbook of Microbial Iron Chelates. G. Winkelmann, Ed. CRC Press, Boca Raton, Fla., USA, pp: 1-14. Haag, H., H.P. Fiedler, J. Meiwes, H. Dreschel, G., Jung, G. and H. Zahner, 1994. Isolation and biological characterization of staphyloferrin B, a compound with siderophore activity from staphylococci. FEMS Microbiol. Lett., 115: 125-130. Smith, M. and J. Neilands, 1984. Rhizobactin, a siderophore from Rhizobium meliloti. J. Plant Nutr., 7: 449-458. Smith, M., J. Shoolery, B. Schwyn and J. Neilands, 1985. Rhizobactin, a structurally novel siderophore form Rhizobium meliloti. J. Am. Chem. Soc., 107: 1739-1743. Thieken, A. and G. Winkelmann, 1992. Rhizoferrin: a complexone type siderophore of Mucorales and Entomophthorales (Zygomycetes). FEMS Microbiol. Lett., 94: 37-42. Schwyn, B. and J.B.. Neilands, 1987. Universal chemical assay for the detection and determination of siderophores. Anal. Biochem., 160: 47-56. Langmyhr, F. and K. Klausen, 1963. Complex formation of iron (III) with chrome azurol S. Anal. Chim. Acta., 29: 149-167. Winkelmann, G., B. Busch, A. Hartmann, G. Kirchhoff, R. Submutha and G. Jung, 1999. Degradation of desferrioxamines by Azospirillum irakense: assignment of metabolites by HPLC/electrospray mass spectrometry. Bio-Metals, 12: 255-264. Castignetti, D. and A. Siddiqui, 1990. The catabolism and heterotrophic nitrification of the siderophore deferrioxamine B. Biol. Metals., 3: 197-203. Harwani, S., A. Roginsky, Y. Vallejo and D. Castignetti, 1997. Further characterization and proposed pathway of deferrioxamine B catabolism. Biometals, 10: 205-213. Pierwola, A., T. Krupinski, P. Zalupski, M. Chiarelli and D. Castignetti, 2004. Degradation Pathway and Generation of Monohydroxamic Acids from the Trihydroxamate Siderophore Deferrioxamine B. Appl. Environ. Microbiol., 70: 831-836.

22. Zaya, N., A. Roginsky, J. Williams and D. Castignetti, 1998. Evidence that a deferrioxamine B degrading enzyme is a serine protease. Can. J. Microbiol., 44: 521-527. 23. Warren, R. and J. Neilands, 1964. Microbial degradation of the ferrichrome compounds. J. Gen. Microbiol., 35: 459-470. 24. Warren, R. and J. Neilands, 1965. Mechanism of microbial catabolism of ferrichrome A. J. Biol. Chem., 240: 2055-2058. 25. Winkelmann, G., K. Schmidtkunz and F. Rainey, 1996. Characterization of a novel Spirillum-like bacterium that degrades ferrioxamine-type siderophores. Bio. Metals, 9: 78-83. 26. Villavicencio, M. and J. Neilands, 1965. An inducible ferrichrome A-degrading peptidase from Pseudomonas FC1. Biochem., 4: 1092-1097. 27. DeAngelis, R., M. Forsyth and D. Castignetti, 1993. The nutritional selectivity of a siderophorecatabolizing bacterium. Bio. Metals., 6: 234-238. 28. Wolin, E., M. Wolin and R.S., Wolfe, 1963. Formation of methane by bacterial extracts. J. Biol. Chem., 238: 2882-2886. 29. Hantke, K., 1987. Selection procedure for deregulated iron transport mutants (fur) in Escherichia coli K12: fur not only affects iron metabolism. Mol. Gen. Genet., 210: 1335-139. 30. Sillen, L. and A.E. Martell, 1964. Stability constants of metal-ion complexes. Part I-Inorganic ligands and Part II-Organic ligands. Special Publication No. 17. The Chemical Society. Burlington House, W1. 31. Sillen, L. and A.E. Martell, 1971. Stability constants of metal-ion complexes. Part I-Inorganic ligands and Part II-Organic ligands. Special Publication No. 25. The Chemical Society. Alden and Mowbray, Ltd. at the Alden Press. Oxford. 32. Dover, L. and C. Ratledge, 1996. Identification of a 29 LDO protein in the envelope of Mycobacterium smegmatis as a purative ferri-exochelin receptor. Microbiology, 142: 1521-1530. 33. Stull, T., K. Mack, J. Haas, J. Smit and A. Smith, 1985. A comparison of techniques for isolation of the outer membrane proteins of Haemophilus influenzae type B. Anal. Biochem., 150: 471-480. 34. Blonder, J., M. Goshe, R. Moore, L. Pasa-Tolic, C. Masselson, M. Lipton and R. Smith, 2002. Enrichment of integral membrane proteins for proteomic analysis using liquied chromatographytandem mass spectrometry. J. Proteome Res., 1: 351-360. 128

Advan. Biol. Res., 1 (3-4): 122-129, 2007

35. Chackraborty, R.N. and B. Clark, 2003. Characterization of a catechole type siderophore and the detection of an outer-membrane receptor protein from Rhizobium leginosarum RL 312. Gen. Meet. Am. Soc. Microbiol. 103rd, 2003. Abstr. K-023. 36. Diarra, L., 2003. Characterization of a Staphylococcus aureus that are defective for iron acquisition and assimilation. Gen. Meet. Am. Soc. Microbiol. 103rd, 2003. Abstr. B-322.

37. Simons, K., R. Morton, D. Mosier, R. Fulton and A. Confer, 1989. Comparison of the Pasteurella haemolytica A1 envelope proteins obtained by two cell disruption methods. J. Clin. Microbiol., 27: 664-667.

129