Biodegradation 15: 289–301, 2004. 289 Ó 2004 Kluwer Academic Publishers. Printed in the Netherlands.

Isolation and growth characteristics of an EDTA-degrading member of the a-subclass of

Hans-Ueli Weilenmann, Barbara Engeli, Margarete Bucheli-Witschel & Thomas Egli* Department of Environmental Microbiology and Molecular Ecotoxicology, Swiss Federal Institute for Environmental Science and Technology (EAWAG), U¨berlandstr. 133, 8600 Du¨bendorf, Switzerland (*author for correspondence: e-mail: [email protected])

Accepted 21 April 2004

Key words: aerobic degradation, bacterial strain DSM 9103, complexing agents, EDTA, Rhizobiazeae group,

Abstract A Gram-negative, ethylenediaminetetraacetic acid (EDTA)-degrading bacterium (deposited at the German Culture Collection as strain DSM 9103) utilising EDTA as the only source of carbon, energy and nitrogen was isolated from a mixed EDTA-degrading population that was originally enriched in a column system from a mixture of activated sludge and soil. Chemotaxonomic analysis of quinones, polar lipids and fatty acids allowed allocation of the isolate to the a-subclass of Proteobacteria. 16S rDNA sequencing and phylogenetic analysis revealed highest similarity to the Mesorhizobium genus followed by the genus. However, the EDTA-degrading strain apparently forms a new branch within the / À1 Mesorhizobia family. Growth of the strain was rather slow not only on EDTA (lmax=0.05 h ) but also on other substrates. Classical substrate utilisation testing in batch culture suggested a quite restricted carbon source spectrum with only lactate, glutamate, and complexing agents chemically related to EDTA (nitri- lotriacetate, iminodiacetate and ethylenediaminedisuccinate) supporting growth. However, when EDTA- limited continuous cultures of strain DSM 9103 were pulsed with fumarate, succinate, glucose or acetate, these substrates were assimilated immediately. Apparently, the strain can use a broader spectrum than indicated by traditional substrate testing techniques. The EDTA species CaEDTA and MgEDTA served as growth substrates of the strain because in the mineral medium employed EDTA was predicted to be mainly present in the form of these two complexes. The bacterium was not able to degrade Fe3+ -complexed EDTA.

Introduction phosphonates and particularly aminopolycarb- oxylic acids are extensively used in industrial In a variety of industrial processes as well as in processes and products (Potthoff-Karl et al. 1996). industrial and household products complexing Within the latter group, ethylenediaminetetra- agents are used to sequester and mask metal ions acetic acid (EDTA) and nitrilotriacetic acid in order to prevent the deleterious effects of (NTA) are the predominant chemicals followed free metal ions. Thus, complexing agents are pro- by diethylenetriaminepentaacetic acid (DPTA) duced and applied in large quantities. Worldwide, and hydroxyethylethylenediaminetriacetic acid the most frequently employed complexing agents (HEDTA). are di- and triphosphates, which are essential Today the quantity of EDTA employed components of laundry detergents (Egli 1988). worldwide amounts to some 103,000 of metric tons In addition, hydroxycarboxylates (e.g., citrate), per year (calculated as H4EDTA for 1993; Ko¨ nen 290

1997) and its fields of application are plentiful, tion. To establish biological treatment processes ranging from photographic industry, as galvanic the availability of suitable microbial strains capa- industry, for textile and paper manufacturing, for ble to degrade EDTA is required. To date, only the decontamination of nuclear power installa- two EDTA-consuming pure bacterial cultures tions, as a component of fertilisers and industrial have been described in detail, namely Agrobacte- cleaners, to additives in cosmetics and food prod- rium radiobacter ATCC 55002 (Lauff et al. 1990) ucts (Potthoff-Karl et al. 1996; Wolf & Gilbert and strain BNC1 (No¨ rtemann 1992). Here we de- 1992). scribe the isolation of a third EDTA-utilising Due to its mainly water-based applications strain and its physiological and taxonomic char- EDTA is normally released into wastewater. acterisation. However, in conventional wastewater treatment plants EDTA is neither biologically degraded nor is it known to sorb onto activated sludge (Alder Material and methods et al. 1990; Gardiner 1976; Kari & Giger 1996; Saunama¨ ki 1995). Up to now, there is merely one Media and culture conditions report on the efficient EDTA elimination of about 80% in an industrial wastewater treatment plant The strain was cultivated at 30 °C in batch or (Kaluza et al. 1998). Also in natural systems continuous culture with a mineral salts medium EDTA has been observed to be rather recalcitrant already employed for the isolation of NTA-utilis- towards chemical or biological degradation lead- ing (Egli et al. 1988). It was supplemented ing to high EDTA concentrations measured in with carbon and nitrogen sources as indicated for surface waters that range from about 10 to the individual experiments and the pH was ad- 100 lglÀ1 ( Houriet 1996; Klopp & Pa¨ tsch 1994; justed to 7.0. The mineral medium contained À1 À1 Ko¨ nen 1997). Only the Fe(III)EDTA complex was 1.0 g l MgSO4 Æ 7H2O, 0.2 g l CaCl2 Æ 2H2O, À1 À1 shown to be photolabile and degradable by sun- 0.13 g l KH2PO4 and 0.615 g l Na2HPO4. light (Kari 1994). On sunny days, this process can Two milliliter of Widdel trace element solution contribute to EDTA elimination in rivers. (Pfennig et al. 1981) and 1 ml of a vitamin solution All in all, EDTA has become a concern in re- (Egli et al. 1988) were added to 1 l of mineral cent years, especially because of its possible con- medium. tribution to the mobilisation of heavy metal ions Continuous cultivation of the EDTA-degrad- deposited in river sediments and soils. In Ger- ing strain was performed in bioreactors (MBR, many, for instance, the government and repre- Switzerland, 2 l working volume and Bioengi- sentatives of industry published a statement neering, Switzerland, 2.5 l working volume), in announcing the intention to reduce EDTA pollu- which the pH was maintained at 7.0 by automatic tion of aquatic systems by 50% (BMU 1991). A addition of H3PO4 (1.0 M). The culture was stirred reduction can be achieved with a number of mea- at 1000 rpm and aerated at a rate of 0.1 vvm. sures such as recycling of EDTA in industrial Batch growth experiments to determine maximum processes or by replacing EDTA by other more specific growth rates and to observe growth with easily degradable complexing agents such as NTA mixtures of complexing agents were carried out in or EDDA (Sykora et al. 2001). In addition, 1 l Erlenmeyer flasks containing initially 0.5 l ethylenediaminedisuccinate (EDDS), a structural medium being aerated by vigorous stirring with a isomer of EDTA of biological origin, has been magnetic fly. proposed to substitute EDTA in some applications (Kovaleva et al. 1992). In case of two processes, Characterisation of the EDTA-degrading bacterial i.e. pulp making and radionuclide decontamina- isolate tion, it has been shown, at least theoretically, that EDDS should be more or less equivalent to EDTA Gram staining, aminopeptidase and oxidase tests in terms of efficacy (Jones & Williams 2001). were performed using standard methods. For Alternatively, also physicochemical and bio- metabolic profiling the API test stripes 20NE, 20E logical techniques for the elimination of EDTA and 50CH (Bio Me´ rieux, France) were employed. from industrial wastewaters have received atten- The GC content, the quinone pattern and the 291 polar lipid and fatty acid composition was analy- NTA and fumarate sed by DSMZ (Germany). Also 16S rDNA Both polycarboxylic acids were analysed by sequencing, sequence comparison and presenta- HPLC using a HPICE-AS1 ion exclusion column tion of the results as phylogenetic tree and simi- (Dionex, USA) following a method described by larity matrix was carried out by DSMZ. Schneider et al. (1988).

N,N0-EDDA Biomass determination N,N0-EDDA was analysed as the Cu(II)-complex by HPLC with a Lichrochart 250-4 100 NH2 col- Biomass was routinely monitored as optical den- umn (Merck) by the method of Klu¨ ner (1996). The sity (OD) at 546 nm with the help of an Uvikon eluent contained 1.0 mM cupric acetate plus 860 spectrophotometer (Kontron, Switzerland). 0.17 M acetic acid, with the pH adjusted to 5.1 As during growth with EDTA insoluble precipi- with NaOH. tates were formed the culture liquid was acidified with a few microliters of concentrated HCl prior to IDA the determination of OD. Additionally, biomass IDA concentrations were measured by HPLC with was quantified gravimetrically as dry weight (DW) a HPIC AS11 anion exchange column (Dionex). by filtration of cell suspension through a 0.2 lm For separation a NaOH gradient was employed pore size polycarbonate membrane filter (Nucle- increasing from initially 2.5% NaOH (95 mM) to pore, USA). Cells collected on filters were washed 7.5% NaOH (95 mM) within 6 min at a flow rate with acidified distilled water and were dried at of 1 ml minÀ1. IDA was detected by a Dionex ° 105 C to constant weight. CD20 conductivity detector.

Analysis of dissolved organic carbon (DOC) and Analysis of nitrogenous compounds substrate concentrations in the cultivation liquid Ammonium concentrations were measured accor- Prior to analysis cells were removed from the ding to the indophenol method described by cultivation liquid either by centrifugation (5 min, Scheiner (1976). Rapid determination of the 20,000 · g) or by filtration through a 0.2 lm pore ammonium concentration in the bioreactor culti- size polyvinylidene difluorid membrane filter vation liquid was performed with analytical test (Millipore, USA) and the cell-free sample was then strips provided by Merck. used for determination of the following para- Nitrate and nitrite were analysed simulta- meters. neously with the help of an automated ion analyser (Skalar, The Netherlands).

DOC Chemicals The filtrate was appropriately diluted with dis- tilled, carbon-free water and the solution was [S,S]-EDDS was a gift of the laboratory of Prof. acidified with one drop of HCl 32%. After strip- H. Za¨ hner (Tu¨ bingen, Germany). All other ping dissolved CO2 with N2 gas DOC was mea- chemicals were of analytical grade and were pur- sured with a TOCOR 2 total organic carbon chased from either Fluka (Switzerland) or Merck. analyser (Maihak AG, Germany).

EDTA and [S,S]-EDDS Results These two compounds were measured as the iron(III)-complex by HPLC with a Lichrochart Isolation of an EDTA-degrading strain 250-4 C18 column (Merck, Germany) as described previously by Witschel & Egli (1998). The eluent Description of the inoculum consisted of a formate buffer (5 mM sodium for- The inoculum for the isolation of an EDTA- mate, 15 mM formic acid, 2 mM tetrabutylam- degrading bacterial strain was obtained monium bromide) and contained 12.5% methanol. from N. Gschwind (Ciba-Geigy, Schweizerhalle, 292

Switzerland). It was a mixed microbial culture As a consequence, the mixed culture grew faster enriched in a column packed with activated carbon and did not flocculate any more. Concomitantly that was continuously fed with a mineral medium with the consumption of EDTA biomass was containing EDTA as sole source of carbon, formed and nitrate accumulated in the medium. nitrogen and energy. The column was originally Only a small increase of ammonium and nitrite inoculated with activated sludge from various concentrations in the culture liquid was observed. industrial wastewater treatment plants and with About 60% of the EDTA-nitrogen were found in soil extracts. In the mixed population selected the form of inorganic nitrogen compounds at the EDTA degradation with concomitant nitrate end of the cultivation period. The accumulation of production was observed (Gschwind 1992). nitrate instead of ammonium suggested the pres- ence of nitrifying bacteria in the mixed commu- + nity, which oxidised the NH4 excreted by the Further enrichment on a column EDTA-degrader. A column packed with glass beads was inoculated with the mixed culture above described. The col- umn was aerated and fed with a mineral salt Isolation of a pure EDTA-degrading strain from medium described in Materials and methods con- suspended batch cultures taining between 200 and 300 mg lÀ1 EDTA as Cells from batch cultures were diluted in phos- only source of carbon and nitrogen. The medium phate buffer (pH 7.2) and plated on Plate Count was circulated and the DOC concentration in the Agar (PCA). From PCA plates single colonies liquid was followed routinely. When it indicated were picked and grown in liquid culture with an that EDTA was nearly consumed fresh EDTA was optimised mineral medium containing EDTA. added to the medium. Direct plating on selective EDTA agar plates was not successful because colony growth was too slow. This procedure was repeated until a pure Enrichment in suspended batch cultures culture was obtained. The pure EDTA-degrading After about half a year of continuous enrichment strain was deposited at the German Culture Col- in the glass bead column system EDTA degrada- lection under the number DSM 9103. tion became stable and reproducible. Material The maximum specific growth rate, lmax,of from the column was then used as inoculum for strain DSM 9103 with EDTA as the only source batch cultures. The same EDTA-containing med- of carbon, nitrogen and energy (initial concen- ium as in the column enrichment was used. Since tration of 850 mg EDTA lÀ1) at pH 7.0 and the cells flocculated strongly in batch culture sev- T=30 °C was 0.05 ± 0.006 hÀ1. EDTA was uti- eral improvements of the culture conditions were lised down to a final concentration below introduced: First the Ca2+ concentration in the 5mglÀ1. Relating measured DOC concentrations medium was doubled. Second, the pH optimum to those of EDTA indicated that no carbonaceous and temperature optimum as well as the optimum products accumulated to a significant extent dur- EDTA concentration for growth were determined. ing growth with EDTA. However, surplus nitro- The mixed culture grew best at slightly basic pH gen stemming from EDTA accumulated in the values (optimum between pH 7.5 and 8.5) and at a culture liquid in the form of ammonium and temperature of 30–32 °C. Initial EDTA concen- reached a final concentration of about 90 mg À1 þ À1 trations of 1–2 g l did not inhibit growth of the NH4 l (Figure 1). The growth yield, Yx/s, for enriched bacterial community. Based on these re- EDTA was 0.26 ± 0.01 g DW (g EDTA)À1. Both sults the following growth conditions for EDTA the specific growth rate and the yield of DSM degraders were chosen for further experiments: 9103 on EDTA were similar to those reported for T=30 °C, initial EDTA concentration in the growth of the EDTA-degrading strain BNC1 un- À1 Æ range of 1–1.5 g l , CaCl2 2H2O concentration der similar conditions, which were lmax=0.03– À1 À1 À1 in the mineral medium of 0.4 g l , and an initial 0.07 h and Yx/s=0.271 g DW (g EDTA) , pH of 7.0. Instead of slightly basic initial pH of the respectively (Henneken et al. 1994, 1998). No medium a neutral pH was chosen because during growth of strain DSM 9103 with EDTA was ob- consumption of EDTA pH increased anyhow. served under denitrifying conditions. 293

900 0.8

800 0.7 700 0.6 600 0.5 500

0.4 546 400 OD

, and DW [mg/l] 0.3 + 300

NH4 200 0.2

Concentrations of EDTA, DOC, 100 0.1 0 0 0204060 Time [h]

Figure 1. Batch growth of strain DSM 9103 with EDTA as sole source of carbon, nitrogen and energy. The biomass concentration in þ the cultivation liquid was monitored as DW () and OD546 (). In addition, the consumption of EDTA (n), the excretion of NH4 (w) as well as the development of DOC (h) were recorded.

Characterisation of strain DSM 9103 growth tests in test tubes using liquid mineral medium amended with various carbon substrates Strain DSM 9103 was a Gram-negative, non-mo- were negative despite extended periods of incu- tile, rod-shaped bacterium. The short rods were bation for up to 2 months. Only lactate and glu- often found as pairs under the microscope (Figure tamate indicated scanty growth. However, slow 2). It formed small white colonies on EDTA but good growth, was observed on complex plates. According to results from the API test medium at low concentrations, such as Plate stripes the strain is urease and oxidase positive Count Broth at one third of the normal strength. and can transform esculin. All other API tests Moreover, pulsing of carbon sources into a con- were negative or not clearly positive. Neither un- tinuous culture of strain DSM 9103 growing with der aerobic nor under fermenting conditions was EDTA as sole source of carbon, nitrogen and growth observed with the carbon sources offered energy (D=0.035 hÀ1) showed that the bacteria in the API 20NE, 20E and 50CH stripes. Also were able to immediately utilise glucose, acetate,

Figure 2. Electron micrograph of thin sections of cells of strain DSM 9103 grown with EDTA as sole source of carbon, nitrogen and energy (the bar represents 1 lm). 294 succinate and fumarate as growth substrates. sequences of members of the a-subclass of Prote- Subsequently, growth with succinate or fumarate obacteria. Using only type strains for the various as sole source of carbon and energy was also ob- genera, the results are presented as phylogenetic served in batch culture when using bigger culti- tree (Figure 3) and similarity matrix (Table 1). vation volumes. Highest similarity was found between the 16S The GC content of the EDTA-degrading iso- rDNA sequences of strain DSM 9103 and mem- late was 63.1 mol%. No menaquinones were found bers of the Mesorhizobium genus with 96.2% sim- in strain DSM 9103 and the major ubiquinone was ilarity between DSM 9103 and the type strain one with 10 isoprenoid units in the side chain Mesorhizobium loti. Furthermore, the 16S rDNA (ubiquinone Q-10). This is indicative of a position sequence of DSM 9103 showed high similarity to of the strain in the a-subclass of Proteobacteria. the 16S rDNA of Aminobacter aminovorans and The polar lipid pattern suggests the strain to be a Chelatobacter heintzii (95.7% similarity) and of member of either the Paracoccus branch of the Pseudoaminobacter salicylatoxidans (95.6% simi- a-subclass, or the Rhizobium and Agrobacterium larity). As C. heintzii was proposed to be a syno- group. Hydroxy fatty acids were not present and nym of Aminobacter aminovorans (Ka¨ mpfer et al. the predominant fatty acids were the ones com- 2002) only the latter species is represented in the prised in ‘summed feature 7’ (18:1 x7c/x9t/x12t phylogenetic tree in Figure 3. However, 16S and 18:1 x7c/x9c/x12t) with a portion of 67.63% rDNA similarities were below the 97% mark re- followed by 19:0 cyclo x8c with 23.05%. quested for the affiliation of a new species to an The complete sequence of the 16S rRNA gene existing genus. Hence, isolate DSM 9103 seems to of DSM 9103 was determined by direct sequencing belong to a new genus within the Rhizobiales order of the PCR amplified 16S rRNA DNA. The se- and, more precisely, within the Mesorhizobium/ quence was compared with the currently available Phyllobacterium group (family).

Figure 3. Dendrogram of type strains of the a-subclass of Proteobacteria most closely related to the EDTA-degrading strain DSM 9103. Since similarities of greater than 99.6% were found between the 16S rDNA sequences of Aminobacter aminovorans and Che- latobacter heintzii (Ka¨ mpfer et al. 1999) and C. heintzii was reported to be a subjective synonym of A. aminovorans (Ka¨ mpfer et al. 2002), only the latter was included in the tree although both strains are still treated as two separate genera in the new edition of Bergey’s manual. 295

Table 1. Percentage similarities between 16S rDNA sequences from type strains of the genera Mesorhizobium, Aminobacter, Pseudoaminobacter and other members of the Rhizobiales order as well as the related EDTA-degrading isolate DSM 9103

Strains 1 2 345678910111213

1. DSM 9103 – 2. Mesorhizobium loti 96.2 – 3. Pseudoaminobacter salicylatoxidans 95.6 95.9 – 4. Phyllobacterium myrsinacearum 94.9 96.6 95.7 – 5. Aquamicrobium defluvii 95.0 95.4 96.5 95.8 – 6. Defluvibacter lusatiensis 95.1 95.3 95.5 95.5 98.4 – 7. Sinorhizobium fredii 94.9 95.1 94.7 96.0 94.9 95.3 – 8. Aminobacter aminovorans 95.7 98.0 96.9 96.6 95.7 95.2 95.0 – 9. Rhizobium leguminasorum bv. trifolii 93.3 93.5 92.9 93.7 93.5 93.9 95.3 93.3 – 10. Ochrobactrum anthropi 93.3 93.8 92.7 94.4 92.9 94.2 94.9 93.7 94.3 – 11. Brucella melitensis 94.3 94.4 93.8 95.5 94.4 95.5 96.0 94.2 94.6 97.4 – 12. Bartonella bacilliformis 93.0 93.8 92.8 95.1 93.8 94.0 94.9 93.5 93.4 93.4 95.0 – 13. Azorhizobium caulinodnas 89.4 89.2 90.0 90.2 90.0 90.5 89.4 89.7 89.7 90.5 90.5 88.4 –

Influence of metal ions on the utilisation of EDTA times lower (Kt=0.4 lM), suggesting that the culture should exhibit a higher affinity than re- Both in nature and in growth media used in this flected by the value obtained for Ks. When using Kt work powerful chelating agents such as EDTA are as affinity constant in the Monod equation a complexed with metal ions. Therefore, the influ- residual EDTA concentration of about 1 lMis ence of metal ion complexation on the utilisation predicted. Considering that the concentration of of EDTA by strain DSM 9103 had to be consid- all heavy metal ions in the medium was 15–20 lM ered. one has to assume that most of the residual EDTA The initial speciation of EDTA in the mineral in the cultivation liquid was complexed with heavy medium employed for the cultivation of strain metal ions and was therefore not available to the DSM 9103 was predicted with the help of the cells. speciation prediction program ChemEQL (Mu¨ ller Since Fe3+ ions make up for the largest por- 1996). Accordingly, EDTA (initial concentration tion of all heavy metal ions in the medium em- of 1.5 g lÀ1) should be mainly complexed with ployed we tested the utilisation of Fe(III)EDTA Mg2+ (58.3% of total EDTA) and Ca2+ (41.1% of by strain DSM 9103 in an additional experiment. total EDTA) (Witschel et al. 1999). This indicates We focused on the degradability of Fe(III)EDTA that strain DSM 9103 is able to utilise EDTA for two further reasons. First, in the environment when complexed with either Mg2+ or with Ca2+ a large portion of EDTA may be complexed with or with both of them. Fe3+ ions (Kari 1994; Nowack 1996). For in- In a chemostat experiment residual EDTA stance, in the influent of several Swiss municipal concentrations were determined when the culture wastewater treatment plants the fraction of grew at D=0.035 hÀ1 and was fed with a mineral Fe(III)EDTA was found to range between 20% medium containing approximately 1.5 g lÀ1 and 50%, in the effluent it was even higher (40– EDTA. The residual concentrations were about 80%), and in the Swiss river Glatt the fraction 36 lM. Using a maximum specific growth rate amounted to 30%. Second, EDTA-utilising Agro- with EDTA of 0.05 hÀ1 a half saturation constant bacterium radiobacter ATCC 55002, was reported Ks of approximately 15 lM was estimated with the to be able to grow with Fe(III)EDTA only (Lauff help of the Monod equation. This contrasts et al. 1990) whereas EDTA-degrading strain strongly with the apparent affinity constant of the BNC1 was found to be unable to use Fe(III)- EDTA transport system reported earlier for complexed EDTA (Henneken et al. 1995). To test CaEDTA (Witschel et al. 1999) which was 30–40 whether strain DSM 9103 can degrade EDTA 296 when added as Fe(III)-complex, Fe(III)EDTA was the ongoing consumption of the inflowing EDTA, pulsed into a continuous culture of strain DSM the formation of biomass was immediately inhib- À1 9103 growing at D=0.025 h and fed with a ited as indicated by the OD546 following the mineral medium containing 1.5 g lÀ1 EDTA as theoretical wash-out curve perfectly. After 4–5 h, sole source of carbon, nitrogen and energy. The EDTA started to accumulate in the culture and initial concentration of the added Fe(III)EDTA wash-in of EDTA proceeded until 24 h after the was approximately 250 mg lÀ1. For comparison, a pulse indicating an inhibition of the consumption pulse of uncomplexed EDTA was added to a of otherwise readily degradable EDTA complexes. continuous culture growing at D=0.015 hÀ1. The phenomenon that, after some time lag, the Considering the excess of Mg2+ and Ca2+ in the presence of Fe(III)EDTA exerts an inhibitory ef- mineral medium one can predict that the pulsed fect on the degradation of free EDTA has also free EDTA was instantaneously complexed by previously been observed in degradation experi- Mg2+ or Ca2+ present in the culture medium. ments with resting cells of DSM 9103 (Satroutdi- The response of the culture was considerably nov et al. 2000). After 24 h the measured EDTA different for the two pulses. In case of concentrations dropped again below the theoreti- Fe(III)EDTA addition (Figure 4) the concentra- cal wash-in curve. Hence, consumption of EDTA tion of EDTA in the cultivation liquid remained from the medium resumed before Fe(III)EDTA more or less constant during the first few hours had been washed out completely finally resulting in after the pulse and then started to increase. A regrowth of biomass. comparison of measured EDTA concentrations In contrast, in case of a pulse of uncomplexed with calculated wash-out and wash-in curves sug- EDTA (data not shown) the comparison of the gested that during the first hours after the pulse the measured EDTA concentrations with the theoret- EDTA continuously supplied to the culture from ical wash-out curve suggested immediate utilisa- the medium reservoir was consumed whereas tion of the added EDTA. During the pulse, the the extra Fe(III)EDTA was not utilised. Despite measured DOC concentrations in the culture

1.2

900

1.0 546 800

700 0.8 600

500 0.6

400 0.4 300 and theoretical wash-out of OD and theoretical wash-out 546 of EDTA from the medium [mg/l] of EDTA from the medium [mg/l] 200 0.2 OD DOC stemming from EDTA [mg C/l] and DOC stemming from EDTA [mg C/l] and 100 Concentrations of EDTA [mg/l], DOC [mg C/l], Concentrations of EDTA [mg/l], DOC [mg C/l], theoretical wash-out of Fe(III)EDTA and wash-in of Fe(III)EDTA and wash-in theoretical wash-out theoretical wash-out of Fe(III)EDTA and wash-in of Fe(III)EDTA and wash-in theoretical wash-out

00 0.0 -5 0 5 10 15 20 25 30 35 Time [h]

Figure 4. Pulse-wise addition of Fe(III)-complexed EDTA to a continuous culture of strain DSM 9103 growing with EDTA as sole source of carbon, nitrogen and energy (D=0.025 hÀ1). The Fe(III)EDTA pulse was introduced at t=0 h. As a function of time the biomass concentration was measured as OD546 () and compared to the theoretical washout curve (---) calculated under the assumption that the bacteria stopped consuming EDTA. Moreover, the EDTA concentration in the cultivation liquid (n) was monitored and based on these values the DOC stemming from EDTA (+) was calculated. The theoretical wash-out curve for the added Fe(III)EDTA (---) and the wash-in curve for EDTA (---) due to a stop of microbial EDTA degradation were also shown. Finally, the DOC (h) was followed during the experiment. 297 always matched those of EDTA and therefore, the fumarate a similar yield was observed. The EDDS consumption of the pulsed EDTA was apparently molecule contains two chiral carbon atoms and, not accompanied by the excretion of any metab- therefore, three optical isomers of EDDS exist, olite. Degradation of EDTA led to the formation namely S,S-EDDS, R,R-EDDS and R,S-EDDS. of additional biomass. Besides S,S-EDDS strain DSM 9103 utilised R,S- EDDS but not R,R-EDDS. Growth with complexing agents other than EDTA Batch growth of strain DSM 9103 with mixtures of Several other complexing agents, all belonging to EDTA and an alternative substrate, in particular a the group of aminopolycarboxylic acids, were as- second aminopolycarboxylic acid sessed as growth substrates for strain DSM 9103 in batch culture in at least two experiments. To obtain some information on the regulation of Diethylenetriaminepentaacetate (DTPA) and EDTA degradation in the presence of other sub- hydroxyethylethylenediaminetriacetate (HEDTA), strates strain DSM 9103 was cultivated under two other important and powerful chelating batch conditions with mixtures of EDTA and agents, did not support growth of strain DSM [S,S]-EDDS, NTA, IDA, N,N0-EDDA and fuma- 9103. On the other hand, nitrilotriacetate (NTA) rate. In all experiments cells were precultured with was utilised by the bacterium. However, growth EDTA during two batch passages before inocu- with NTA was rather poor with respect to both the lating them into the medium containing the sub- yield Yx/s and to the maximum specific growth rate strate mixture (initial concentration of each À1 lmax (Table 2). The known metabolites of EDTA substrate was about 0.5 g l ). and NTA breakdown, i.e. N,N0-EDDA and IDA, respectively, served also as growth substrates. IDA Growth with EDTA/NTA supported even rather fast growth of the strain, During cultivation of strain DSM 9103 with the comparable with growth on the non-complexing EDTA/NTA mixture a sequential utilisation pat- agent fumarate (Table 2). N,N-EDDA a stereo- tern was observed with EDTA being used prefer- isomer of N,N0-EDDA on the other hand was not entially (Figure 5A). Nevertheless, in the initial utilised by the strain. Eventually, strain DSM 9103 growth phase with EDTA also approximately grew with S,S-EDDS, a structural isomer of 100 mg lÀ1 NTA were utilised at a rate of 6.6 mg EDTA. Growth with S,S-EDDS was slightly faster NTA (g DW)À1 hÀ1. As soon as EDTA was close than with EDTA and resulted in the highest to exhaustion NTA utilisation proceeded at a amount of biomass produced per gram of com- higher rate of about 18.4 mg NTA (g DW)À1 hÀ1. plexing agent used. Only during growth with Most of the biomass was produced from EDTA while consumption of NTA resulted only in a very low increase in biomass concentration with a

Table 2. Maximum specific growth rates lmax and yields Yx/s growth yield on NTA comparable to that found during batch growth of strain DSM 9103 with different carbon during growth of strain DSM 9103 with NTA and nitrogen sources only. The maximum specific growth rate on EDTA À1 À1 in the presence of NTA (lmax=0.033 h ) was Growth substrate lmax (h ) YX/S (g DW À1 significantly lower than that observed for growth (g substrate) ) À1 with EDTA only (lmax=0.05–0.06 h ). This EDTA 0.05 ± 0.006 0.26 ± 0.01 indicates a certain inhibitory effect of NTA exerted [S,S]-EDDS 0.067 ± 0.001 0.39 ± 0.07 on growth with EDTA although NTA did not NTA 0.0057 ± 0.0 0.11 ± 0.02 repress the utilisation of EDTA during mixed IDA 0.094 ± 0.004 0.31 ± 0.0 substrate cultivation. N,N0-EDDA 0.059 ± 0.003 0.29 ± 0.01 þ 0 Fumarate + NH4 0.105 ± 0.02 0.38 ± 0.0 Growth with EDTA/N,N -EDDA Also in this case, EDTA was used first and only Initial concentrations of the carbon/energy/(nitrogen) substrate À1 þ À1 shortly before all EDTA had been consumed were approximately 1.0 g l and that of NH4 was 0.5 g l . 0 All values are averages of at least two independent growth N,N -EDDA started to be degraded, initially with experiments. a comparably high rate, which then slowed down 298

450 0.5 600 0.6 (A) (B) 400 0.45 500 0.5 350 0.4 0.35 300 400 0.4 0.3

250 546 0.25 546 300 0.3 200 OD 0.2 OD C and DOC [mg C/l] 150 - 200 0.2

and DOC [mg C/l] 0.15 100 0.1 EDDA 100 0.1

50 0.05 N,N'- Concentrations of EDTA-C, Concentrations of EDTA-C, NTA-C Concentrations of EDTA-C, 0 0 0 0.0 0 100 200 300 400 500 0 100 200 300 400 500 600 700 Time [h] Time [h]

350 0.9 (C) 0.8 300 0.7 250 0.6 200 0.5 546 150 0.4 OD 0.3

and DOC [mg C/l] 100 0.2 50 0.1 Concentrations of EDTA-C, IDA-C Concentrations of EDTA-C, 0 0.0 0 100 200 300 Time [h]

Figure 5. Growth of bacterial strain DSM 9103 with EDTA and a second substrate in batch culture. The biomass concentration, measured as OD546 (), as well as the concentrations of EDTA-carbon (j), carbon from the second substrate (m) and DOC are shown (h). (A) second substrate=NTA, (B) second substrate=N,N0-EDDA, and (C) second substrate=IDA. significantly (Figure 5B). As in case of growth with Growth with EDTA/IDA EDTA/NTA most of the biomass produced The growth of the bacterial isolate DSM 9103 with throughout the whole cultivation period stemmed a mixture of EDTA and IDA followed a diauxic from the utilisation of EDTA and the maximum pattern with IDA being the preferred substrate specific growth rate lmax during EDTA utilisation (Figure 5C). Nevertheless, a small fraction (about (0.03 hÀ1) was reduced in comparison to growth 10%) of EDTA was already degraded during the with EDTA only. A DOC balance considering phase of IDA utilisation. After growth with IDA EDTA and N,N0-EDDA suggests that during which supported a maximum specific growth rate 0 À1 N,N -EDDA utilisation a product was formed and lmax of 0.08 h the culture entered a lag phase excreted into the medium. Consequently, biomass of approximately 3 days before degradation of formation was very low during the slow degrada- EDTA started and resulted in a second growth 0 À1 tion of N,N -EDDA. phase characterised by a lmax of 0.06 h .

Growth with EDTA/[S,S]-EDDS During batch growth with this mixture both sub- Growth with EDTA/fumarate strates were used and completed simultaneously These two substrates were consumed simulta- supporting a maximum specific growth rate lmax neously during batch cultivation and growth pro- of 0.04 hÀ1 during the exponential phase. ceeded at a maximum specific growth rate of 299

0.075 hÀ1. However, when fumarate was ex- species tested (free EDTA, NiEDTA, CuEDTA, hausted, the utilisation of EDTA still present in Co(III)EDTA) supported growth (Palumbo et al. the growth medium stopped as well, Therefore, 1994). In case of DSM 9103 and BNC1 it has not only about 50–60% of the initial EDTA were uti- been unravelled so far which EDTA species is/are lised within the time period analysed. Hence, really the growth substrates but the data available fumarate had a positive effect on the degradation suggests that CaEDTA and other EDTA species of EDTA but on the other hand EDTA break- with a rather low stability constant are the most down seemed to become dependent on the pres- likely species. First, the increase of the Ca2+ ence of fumarate. This might indicate the necessity concentration in the mineral medium for DSM for reorganisation of the EDTA metabolism when 9103 improved growth of the strain markedly. growing with EDTA alone after growth with Second, experiments with resting cells of strain EDTA in the presence of an alternative, easily BNC1 and DSM 9103 showed that EDTA com- degradable substrate. As far as the catabolic plexed with Ca2+,Mg2+,Ba2+,orMn2+ were pathway is known for this complexing agent in readily degraded (Klu¨ ner et al. 1998; Satroutdinov isolate DSM 9103 (Witschel 1999), it includes the et al. 2000). All these complexes have a stability initial investment of electrons and involvement of constant lower than 1016. For more stable EDTA oxygen as co-substrate. One can speculate that complexes (complexes with Co2+,Cu2+,Zn2+ cellular metabolism has to be reorganised because and Pb2+) the results indicated no direct utilisa- in presence of fumarate the electrons needed for tion but dissociation of the complexes prior to substrate activation originate from the degrada- degradation (Satroutdinov et al. 2000). Third, tion of fumarate while during growth with EDTA uptake studies in strain DSM 9103 with 45Ca2+ in alone they may come from the oxidation of the presence and absence of EDTA demonstrated downstream intermediates. that Ca2+ is transported together with EDTA into the cells (Witschel et al. 1999). In the uptake assays and the experiments with resting cells also free EDTA was transported/used. However, strain Discussion BNC1 did not grow with free EDTA and cell lysis caused by an excess of the uncomplexed chelating So far, only two pure EDTA-degrading strains agent was suggested as the reason (No¨ rtemann have been described in detail. Mainly based on 1992). physiological and biochemical tests one strain The proposed close taxonomic relationship of (ATCC 55002) was identified as Agrobacterium strain BNC1 and DSM 9103 is also supported by radiobacter sp. (Lauff et al. 1990; Palumbo et al. the fact that EDTA metabolism in both strains 1994), now renamed Rhizobium radiobacter seems to be very similar if not identical. While (Young et al. 2001), whereas the data collected for EDTA transport has only been investigated in the second strain (BNC1) allowed only its affilia- strain DSM 9103 (Witschel et al. 1999) the initial tion to the a-subclass of Proteobacteria (No¨ rte- intracellular enzymatic steps in EDTA degrada- mann 1992). Here, we present the isolation of a tion were elucidated in both strains (Bohuslavek third EDTA-degrading bacterial strain and some et al. 2001; Witschel et al. 1997): A monooxy- of its physiological, biochemical, chemotaxonomic genase consuming FMNH2 delivered by a and phylogenetic characteristics. NADH2:FMN oxidoreductase catalyses the oxi- The new isolate (strain DSM 9103) has several dation of EDTA and the first intermediate of features in common with strain BNC1, for in- EDTA, ethylenediaminetriacetate (ED3A). Based stance slow growth on both EDTA and other on SDS PAGE similar molecular weights were carbon sources and the fact that iron-complexed estimated for the enzymes from both isolates and EDTA supports no growth. Both isolates are finally the 26 N-terminal amino acids of the clearly distinct from the third strain, R. radiobacter EDTA monooxygenases of both strains were sp., in particular with respect to the EDTA species identical. used as growth substrate. In contrast to strain Eventually, 16S rRNA sequence analysis re- BNC1 and DSM 9103 R. radiobacter is only able vealed a rather high similarity of DSM 9103 with to grow with Fe(III)EDTA. No other EDTA C. heintzii. The latter strain was originally isolated 300 for its ability to degrade NTA (Auling et al. 1993; Acknowledgements Egli et al. 1988) and it grows also with the ami- nopolycarboxylic acid IDA but not with EDTA or We are indebted to D. Gschwind (Ciba-Geigy, EDDS (Bucheli-Witschel & Egli 2001). Interest- Schweizerhalle, Switzerland) for kindly supplying ingly, EDTA catabolism in strains DSM 9103 and us with the EDTA-degrading mixed microbial BNC1 as well as the enzymes participating in consortium. We further thank Prof H. Za¨ hner EDTA degradation share many common charac- (University of Tu¨ bingen, Germany) for the gen- teristics with those found for the metabolism of erous gift of S,S-EDDS. Our thanks also go to E. NTA in C. heintzii (Egli 1994; Witschel et al. Wehrli (ETH, Zu¨ rich, Switzerland) for the prepa- 1997). The sequence of the gene encoding EDTA ration of the electron micrographs of strain DSM monooxygenase in strain BNC1 was elucidated 9103. This work was financially supported by a and showed the highest similarity/identity values grant from BUWAL. (46.2%/37.8%) when compared with the NTA monooxygenase gene from C. heintzii (Bohuslavek et al. 2001). An interesting observation is the fact that iso- References late DSM 9103 appeared in traditional substrate tests as a rather fastidious organism with an ex- Alder AC, Siegrist H, Gujer W & Giger W (1990) Behaviour of NTA and EDTA in biological wastewater treatment. Water tremely restricted spectrum of carbon substrates Res. 24: 733–742 that can be used. However, when cultivated in Auling G, Busse H-J, Egli T, El-Banna T & Stackebrandt E EDTA-limited continuous culture DSM 9103 was (1993) Description of the gram-negative, obligately aerobic, able to utilise and grow with a number of such nitrilotriacetate (NTA)-utilizing bacteria Chelatobacter hei- ‘non-growth’ substrates when these were added as ntzii, gen. nov., sp. nov., and Chelatococcus asaccharovorans, gen. nov., sp. nov. Syst. Appl. Microbiol. 16: 104–112 pulses. This suggests that the substrate utilisation BMU: Bundesminister fu¨ r Umwelt, Naturschutz und Reak- spectra determined for isolates in the traditional torsicherheit, Germany (1991) Erkla¨ rung zur Reduzierung way may not reflect the whole catabolic spectrum der Gewa¨ sserbelastung durch EDTA of an organism, in particular in case of slow Bohuslavek J, Payne J, Liu Y, Bolton H & Xun L (2001) growing bacteria. Thus, the extensive lists that can Cloning, sequencing, and characterization of a gene cluster involved in EDTA degradation from the bacterium BNC1. presently be found in Bergey’s manual and similar Appl. Environ. Microbiol. 67: 688–695 taxonomy reference books may only reflect a Bucheli-Witschel M & Egli T (2001) Environmental fate and fraction of the substrates that can be utilised by microbial degradation of aminopolycarboxylic acids. FEMS the genospecies listed. Microbiol. Rev. 25: 69–106 Finally, batch cultivation of bacterium DSM Egli T (1988) (An)aerobic breakdown of chelating agents used in household detergents. Microbiol. Sci. 5: 36–41 9103 with substrate mixtures gave first indications Egli T (1994) Biochemistry and physiology of the degradation on the regulation of EDTA breakdown, a point of of nitrilotriacetic acid and other complexing agents. In: high importance if the strain is to be employed for Ratledge C (Ed) Biochemistry of Microbial Degradation (pp the treatment of EDTA-containing wastewaters. 179–195). Kluwer Academic Publishers, Dordrecht Alternative substrates supporting a specific Egli T, Weilenmann HU, El-Banna T & Auling G (1988) Gram- negative aerobic, nitrilotriacetate-utilizing bacteria from growth rate of strain DSM 9103 comparable to wastewater and soil. Syst. Appl. Microbiol. 10: 297–305 that with EDTA apparently do not repress EDTA Gardiner J (1976) Complexation of trace metals by ethylenedi- utilisation totally when present in the medium. aminetetraacetic acid (EDTA) in natural waters. Water Res. However, EDTA consumption stopped when 10: 507–514 these substrates were used up suggesting the Gschwind N (1992) Biologischer Abbau von EDTA in einem Modellabwasser. gwf Wasser Abwasser 133: 546–549 necessity of an intracellular rearrangement to de- Henneken L, Klu¨ ner T, No¨ rtemann B & Hempel DC (1994) grade EDTA only. This points to a rather low Abbau von EDTA mit freien und immobilisierten Bakterien. metabolic flexibility of strain DSM 9103, a gwf Wasser Abwasser 135: 354–358 property that might be a disadvantage for the Henneken L, No¨ rtemann B & Hempel DC (1995) Influence of application of DSM 9103 for industrial waste- physiological conditions on EDTA degradation. Appl. Microbiol. Biotechnol. 44: 190–197 water treatment processes because fluctuations in Henneken L, No¨ rtemann B & Hempel DC (1998) Biological the composition of wastewater are the rule rather degradation of EDTA: Reaction kinetics and technical than the exception. approach. J. Chem. Technol. Biotechnol. 73: 144–152 301

Houriet J-P (1996) NTA dans les eaux. Cahier de l’environn- Palumbo AV, Lee SY & Borman P (1994) The effect of media ement 264. Office fe´ de´ ral de l’environnement, des foreˆts et du composition on EDTA degradation by Agrobacterium sp. paysage (OFEFP), Berne, Switzerland Appl. Biochem. Biotechnol. 45/46: 811–822 Jones PW & Williams DR (2001) Speciation efficiency indices Pfennig N, Widdel F & Tru¨ per HG (1981) The dissimilatory (SEI) and readily-biodegradable indices (RBI) for optimising sulfate-reducing bacteria. In: Starr MP, Stolp H, Tru¨ per HG, ligand control of environmental and associated industrial Balows H & Schlegel HG (Eds) The Prokaryotes, Vol 1 (p processes. Int. J. Environ. Anal. Chem. 81: 73–88 931). Springer Verlag, Berlin Kaluza U, Klingelho¨ fer P & Taeger K (1998) Microbial Potthoff-Karl B, Greindl T & Oftring A (1996) Synthese degradation of EDTA in an industrial wastewater treatment abbaubarer Komplexbildner und ihre Anwendung in Was- plant. Water Res. 32: 2843–2845 chmittel- und Reinigungsformulierungen. SO¨ FW-J. 122: 392– Ka¨ mpfer P, Mu¨ ller C, Mau M, Neef A, Auling G, Busse H-J, 397 Osborn AM & Stolz A (1999) Description of Pseudoaminob- Satroutdinov AD, Dedyukhina EG, Chistyakova TI, Witschel acter gen. nov. with two new species, Pseudoaminobacter M, Minkevich IG, Eroshin VK & Egli T (2000) Degradation salicylatoxidans sp. nov. and Pseudoaminobacter defluvii sp. of metal EDTA complexes by resting cells of the bacterial nov. Int. J. Syst. Bacteriol. 49: 887–897 strain DSM 9103. Environ. Sci. Technol. 34: 1715–1720 Ka¨ mpfer P, Neef A, Salkinoja-Salonen MS & Busse H-J (2002) Saunama¨ ki R (1995) Treatability of wastewaters from totally Chelatobacter heintzii is a later subjective synonym of Amin- chlorine-free bleaching. Tappi J. 78: 185–192 obacter aminovorans. Int. J. Syst. Evol. Microbiol. 52: 835–839 Scheiner D (1976) Determination of ammonia and Kjeldahl Kari FG (1994) Umweltverhalten von Ethylendiamintetraace- nitrogen by indophenol method. Water Res. 10: 31–36 tat (EDTA) unter spezieller Beru¨ cksichtigung des photo- Schneider RP, Zu¨ rcher F, Egli T & Hamer G (1988) Determi- chemischen Abbaus. PhD thesis No. 10698, Swiss Federal nation of nitrilotriacetate in biological matrices using ion Institute of Technology, Zu¨ rich, Switzerland exclusion chromatography. Anal. Biochem. 173: 278–284 Kari FG & Giger W (1996) Speciation and fate of ethylene- Sykora V, Pitter P, Bittnerova I & Lederer T (2001) Biode- diaminetetraacetate (EDTA) in municipal wastewater treat- gradability of ethylenediamine-based complexing agents. ment. Water Res. 30: 122–134 Water Res. 35: 2010–2016 Klopp R & Pa¨ tsch B (1994) Organische Komplexbildner in Uetz TA (1992) Biochemistry of nitrilotriacetate degradation Abwasser, Oberfla¨ chenwasser und Trinkwasser, dargestellt in obligately aerobic, Gram-negative bacteria. PhD thesis am Beispiel der Ruhr. Wasser & Boden: 32–37 No. 9722, Swiss Federal Institute of Technology, Zu¨ rich, Klu¨ ner T (1996) Chemie und Biochemie des mikrobiellen Switzerland EDTA-Abbaus. PhD thesis, Universita¨ t-Gesmthochschule Wanner U, Kemmler J, Weilenmann H-U, Egli T, El-Banna T Paderborn, Germany & Auling G (1990) Isolation and growth of a bacterium able Klu¨ ner T, Hempel DC & No¨ rtemann B (1998) Metabolism of to degrade nitrilotriacetic acid under denitrifying conditions. EDTA and its metal chelates by whole cells and cell-free Biodegradation 1: 31–41 extracts of strain BNC1. Appl. Microbiol. Biotechnol. 49: Witschel M (1999) Biochemical and physiological characterisa- 194–201 tion of a bacterial isolate able to grow with EDTA and other Knobel H-R (1997) Genetic study of bacterial nitrilotriacetate aminopolycarboxylic acids. PhD thesis No. 12967, Swiss degrading enzymes. PhD thesis No. 12146, Swiss Federal Federal Institute of Technology, Zu¨ rich, Switzerland Institute of Technology, Zu¨ rich, Switzerland Witschel M & Egli T (1998) Purification and characterization of Ko¨ nen I (1997) Bestimmung von EDTA-Ersatzstoffen auf a lyase from the EDTA-degrading bacterial strain DSM 9103 Aminopolycarbonsa¨ urebasis. PhD thesis, TH Aachen, Ger- that catalyzes the splitting of [S,S]-ethylenediaminedisucci- many nate, a structural isomer of EDTA. Biodegradation 8: 419– Kovaleva IB, Mitrofanova ND & Martynenko LI (1992) The 428 structure of transition metal complexonates derived from Witschel M, Egli T, Zehnder AJB, Wehrli E & Spycher M ethylenediamine-(N,N0)-disuccinic acid. Russ. J. Inorg. (1999) Transport of EDTA into cells of the EDTA-degrading Chem. 37: 41–45 strain DSM 9103. Microbiology 154: 973–983 Lauff JJ, Steele DB, Coogan LA & Breitfeller JM (1990) Witschel M, Nagel S & Egli T (1997) Identification and Degradation of the ferric chelate of EDTA by a pure culture characterization of the two-enzyme system catalyzing the of an Agrobacterium sp. Appl. Environ. Microbiol. 56: 3346– oxidation of EDTA in the EDTA-degrading bacterial strain 3353 DSM 9103. J. Bacteriol. 179: 6937–6943 Mu¨ ller B (1996) ChemEQL: A program to calculate chemical Wolf K & Gilbert PA (1992) EDTA-Ethylenediaminetetraace- speciation equilibria, titrations, dissolutions, precipitation, tic acid. In: Hutzinger O (Ed) The Handbook of Environ- adsorption, simple kinetics, and pX–pY diagrams. Kasta- mental Chemistry, Vol 3 (pp 243–259). Springer, Berlin nienbaum, Switzerland Heidelberg No¨ rtemann B (1992) Total degradation of EDTA by mixed Young JM, Kuykendall LD, Martinez-Romero E, Kerr A & cultures and a bacterial isolate. Appl. Environ. Microbiol. 58: Sawada H (2001) A revision of Rhizobium, with an emended 671–676 description of the genus, and the inclusion of all species of Nowack B (1996) Behaviour of EDTA in groundwater – a Agrobacterium and Allorhizobiuzm undicola as new combina- study of the surface reactions of metal-EDTA complexes. tions: Rhizobium radiobacter, R. rhizogenes, R. rubi, R. PhD thesis No. 11392, Swiss Federal Institute of Technology, undicola and R. vitis. Int. J. Syst. Evol. Microbiol. 51: 89– Zu¨ rich, Switzerland 103