APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 2010, p. 3220–3227 Vol. 76, No. 100099-2240/10/$12.00 doi:10.1128/AEM.00408-10Copyright © 2010, American Society for Microbiology. All Rights Reserved.

Nonmagnetotactic Multicellular Prokaryotes from Low-Saline,Nonmarine Aquatic Environments and Their Unusual

Negative Phototactic Behavior�†Christopher T. Lefevre,1 Fernanda Abreu,2 Ulysses Lins,2 and Dennis A. Bazylinski1*

School of Life Sciences, University of Nevada at Las Vegas, Las Vegas, Nevada 89154-4004,1 and Instituto de MicrobiologiaProfessor Paulo de Goes, Universidade Federal do Rio de Janeiro, 21941-590 Rio de Janeiro, RJ, Brazil2

Received 13 February 2010/Accepted 23 March 2010

Magnetotactic multicellular prokaryotes (MMPs) are unique magnetotactic bacteria of the Deltaproteobac-teria class and the first found to biomineralize the magnetic mineral greigite (Fe3S4). Thus far they have beenreported only from marine habitats. We questioned whether MMPs exist in low-saline, nonmarine environ-ments. MMPs were observed in samples from shallow springs in the Great Boiling Springs geothermal fieldand Pyramid Lake, both located in northwestern Nevada. The temperature at all sites was ambient, andsalinities ranged from 5 to 11 ppt. These MMPs were not magnetotactic and did not contain magnetosomes(called nMMPs here). nMMPs ranged from 7 to 11 �m in diameter, were composed of about 40 to 60Gram-negative cells, and were motile by numerous flagella that covered each cell on one side, characteristicssimilar to those of MMPs. 16S rRNA gene sequences of nMMPs show that they form a separate phylogeneticbranch within the MMP group in the Deltaproteobacteria class, probably representing a single species. nMMPsexhibited a negative phototactic behavior to white light and to wavelengths of <480 nm (blue). We devised a“light racetrack” to exploit this behavior, which was used to photoconcentrate nMMPs for specific purposes(e.g., DNA extraction) even though their numbers were low in the sample. Our results show that the uniquemorphology of the MMP is not restricted to marine and magnetotactic prokaryotes. Discovery of nonmagne-totactic forms of the MMP might support the hypothesis that acquisition of the magnetosome genes involveshorizontal gene transfer. To our knowledge, this is the first report of phototaxis in bacteria of the Deltapro-teobacteria class.

One of the most interesting and unusual examples of pro-karyotic morphology is that of the organism known as themagnetotactic multicellular prokaryote (MMP; also known asthe magnetotactic multicellular aggregate [MMA] [16, 37] andthe magnetotactic multicellular organism [MMO] [30]). Theacronym MMP originally stood for many-celled magnetotacticprokaryote (45, 46, 53, 54), but in more recent reports, becauseof a number of recent findings suggesting that individual cellsinteract and/or communicate with each other, many investiga-tors use MMP for multicellular magnetotactic prokaryote (foran example, see reference 59).

The MMP is a large prokaryotic microorganism that rangesfrom about 3 to 12 �m in diameter (5, 45, 46). It is bestdescribed as an aggregation of about 10 to 40 (5, 45, 46)Gram-negative, genetically similar cells (53) that swim only asan intact unit and not as individual cells (5, 45, 46). Cells thatseparate from the intact unit die quickly according to viabilitystudies (1). The surface of the cell exposed to the surroundingenvironment is covered with numerous flagella (44, 46, 52).Most described MMPs are spherical (2, 5, 30, 33, 45, 46, 59),although some are ovoid in morphology (36), and all appear to

possess a central, acellular compartment (30, 33). The MMPdivides as aggregates without an individual cell stage (32, 33).In all known instances, MMPs have been found to be magne-totactic, and iron sulfide minerals, magnetic greigite (Fe3S4)and others, in magnetotactic bacteria were first discovered inthis microorganism (17, 39, 43, 44), although magnetite,Fe3O4, has also been identified in some types (31, 38).

MMPs appear to be cosmopolitan in distribution and havebeen observed to be present on most continents in numeroussaline aquatic environments, ranging from brackish (lowestsalinity reported, �12 ppt) (5) to hypersaline (highest salinityreported, �60 ppt) (2, 30, 32, 40). In all cases, the salinity isdue to the input of seawater, and many consider these organ-isms obligately marine (53). More specifically, MMPs havebeen found in stratified sediments of salt marshes (5, 12, 44,53) and coastal lagoons (15–17, 40), the water column ofmeromictic salt water lagoons (44, 54), and some oceanic sed-iments (e.g., Santa Barbara Basin [D. A. Bazylinski, unpub-lished data]) and coastal tidal salt flat sediments (59). Ingeneral, these organisms, like other greigite-producing magne-totactic bacteria, are found in their highest numbers in theanoxic zone of these environments where sulfide is noticeablypresent (5, 53, 54).

Phylogenetically, MMPs are members of the Deltaproteobac-teria, where they form a single cluster within the sulfate-reduc-ing bacteria (2, 12, 53). Simmons and Edwards (53) sequenced16S rRNA genes from a natural population of MMPs from theLittle Sippewissett salt marsh (Falmouth, MA) and found that

* Corresponding author. Mailing address: School of Life Sciences,University of Nevada at Las Vegas, 4505 Maryland Parkway, LasVegas, NV 89154-4004. Phone: (702) 895-2053. Fax: (702) 895-3956.E-mail: dennis.bazylinski@unlv.edu.

† Supplemental material for this article may be found at http://aem.asm.org/.

� Published ahead of print on 2 April 2010.

3220

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

MMPs from this population were phylogenetically made up offive lineages separated by at least 5% sequence divergence.Because of their apparently unique morphology, the MMP wasconsidered by many to consist of a single species, but thisphylogenetic diversity shows that the MMP probably repre-sents a separate genus in the Deltaproteobacteria that includesseveral species. All cells in each aggregate expressed identicalsmall-subunit rRNAs, suggesting that the aggregates are com-posed of a single MMP phylotype (53). Based on the informa-tion presented above, the biogeographical distribution andphylogenetic diversity of the MMP appears to be great, al-though not completely known.

While examining a number of mud and water samples takenfrom low-saline, nonmarine aquatic environments for greigite-producing magnetotactic bacteria, we observed MMP-like or-ganisms. Surprisingly, none of these organisms were magneto-tactic (we refer to them here as nMMPs), a finding notpreviously reported. This is also the first description of theMMP-like morphological form from a nonmarine habitat. Wealso noted that these nMMPs demonstrated an unusual type ofnegative phototaxis that could be exploited to purify enough ofthem for phylogenetic analysis and electron microscopy. In thispaper, we describe these features in detail.

MATERIALS AND METHODS

Sampling sites. In this study, water and sediment samples were taken fromPyramid Lake and several shallow temperate pools in the Great Boiling Springs(GBS) geothermal field in Gerlach, Nevada. Pyramid Lake, one of the largestlakes in the United States, is an endorheic salt lake that covers approximately 490km2 in area. It is located in the Great Basin in the northwestern part of Nevada.GBS is a series of hot springs that range from ambient temperatures to �96°C(4, 11). The geology, chemistry, and microbial ecology of the springs have beendescribed in some detail (4, 11). Pyramid Lake was connected with Gerlachbetween 18,000 and 7,000 BC, during the last ice age, by prehistoric LakeLahontan, which also extended into northeastern California and southern Ore-gon (7, 29). The temperate pools sampled at GBS are designated GL1, GL2, andGL3, and the water temperature in these pools was ambient. GL2 and GL3 arepools about 3 m in diameter formed by old hot springs that no longer have aconnection with geothermally heated water, while GL1 is a shallow marsh closeto an artificial pond used for the introduction of fish to control the breeding ofmosquitoes. For a comparison, we also collected samples containing magneticMMPs from the Salton Sea, which is a saline, endorheic rift lake located on theSan Andreas Fault in southern California. Samples from the Salton Sea weretaken from the southeast shore where the salinity was 52 ppt. Salinities of thesamples were determined with a hand-held Palm Abbe PA203 digital refractom-eter (MISCO Refractometer, Cleveland, OH).

One-liter glass bottles were filled to about 0.2 to 0.3 of their volume withsediment, and the remainder of the bottles were filled to their capacity with waterthat overlaid the sediment in the pool. Air bubbles were excluded. Once in thelaboratory, samples were stored in the dark at room temperature (�25°C). Weobserved nMMPs in these samples over a period of several months. MMPs werenever found in any of these samples at any time, although single-celled magne-totactic bacteria were always present.

Light and electron microscopy. The presence and behavior of microorganismswas observed using light microscopy with a Zeiss (Carl Zeiss MicroImaging, Inc.,Thornwood, NY) AxioImager M1 light microscope equipped with fluorescence,phase-contrast, and differential interference contrast capabilities. The hangingdrop technique (49) was used routinely in the examination of samples and wasalso employed in phototactic behavioral experiments as described below. Inexperiments where numbers of nMMPs required accurate counting, those thataccumulated at the “dark” edge of the drop (the farthest edge of the drop fromthe light beam) were quickly exposed to fluorescent light in the microscope lightusing a Lumar 49 filter set (the same filter set used for cells stained with4�,6-diamidino-2-phenylindole [DAPI]; emission BP445/50, excitation G 365;Carl Zeiss MicroImaging, Inc., Thornwood, NY) which permanently immobi-lized them without disaggregation and facilitated an accurate nMMP count.Because of their large size and conspicuous morphology, nMMPs were easily

distinguishable from single bacterial cells or sediment particles. The presence ofmagnetosomes was determined using electron microscopy with a Tecnai (FEICompany, Hillsboro, OR) model G2 F30 Super-Twin transmission electronmicroscope.

To determine which visible wavelengths of light caused the negative photo-tactic response in the nMMPs, a hanging drop of 5 �l was prepared and themicroscope was focused on one side of the drop. Colored filters were tested withwavelengths of 435 (violet), 480 (blue), 530 (green), and 640 (orange-red) nm(Carl Zeiss MicroImaging, Inc., Thornwood, NY). After 5 min of exposure oflight of a specific wavelength at one edge of the drop, nMMPs that swam to andaccumulated at the opposite edge of the drop were counted as described above.

Determination of 16S rRNA gene sequences and phylogenies of nMMPs. 16SrRNA genes of photoconcentrated nMMPs were amplified using bacterium-specific primers 27F (5�-AGAGTTTGATCMTGGCTCAG-3�) and 1492R (5�-TACGGHTACCTTGTTACGACTT-3�) (35). PCR products were cloned intopGEM-T Easy Vector (Promega Corporation, Madison, WI) and sequenced(Functional Biosciences, Inc., Madison, WI).

Alignment of 16S rRNA genes was performed using the ClustalW multiplealignment accessory application in the BioEdit sequence alignment editor (24).Phylogenetic trees were constructed using Mega version 4 (57), applying theneighbor-joining method (47). Bootstrap values were calculated with 1,000 rep-licates.

Nucleotide sequence accession numbers. 16S rRNA gene sequences ofnMMPs from GBS pools GL1 and GL3 were deposited under GenBank acces-sion numbers GU732821 to GU732823 and GU732824 and GU732825, respec-tively. Those sequences from nMMPs from Pyramid Lake are under GenBankaccession numbers GU732826 and GU732827. The 16S rRNA gene sequence ofthe MMP from the Salton Sea is listed under accession number GU784824.

RESULTS

Description of sampling sites and samples. In all the loca-tions sampled, sediments were black and reducing; the odor ofhydrogen sulfide was readily apparent when sediments weredisturbed. Photosynthetic sulfide-oxidizing bacteria were ob-served macroscopically as patches of pink and microscopicallyin some of the samples that morphologically resembled Chro-matium species.

The salinities of our samples were low compared to those ofseawater (�35 ppt) and other sites where MMPs have beenfound (Pyramid Lake, 6 ppt; GL1, 5 ppt; GL2, 8 ppt; and GL3,11 ppt). There is no marine input to the sampling sites. Despitethese low salinities, these values are still higher than thoseaccepted to be freshwater (�0.5 ppt), and thus the water at thesites must be considered brackish.

Microscopic observations and ultrastructure. One of themost sensitive methods of detecting the presence of magneto-tactic bacteria in natural samples is through the use of thehanging drop technique (49). In addition, the numbers ofnorth- and south-seeking magnetotactic cells can be readilydetermined. While we were examining samples taken fromPyramid Lake and pools and hot springs from the GBS thermalfield in Nevada using the hanging drop technique, we observedrelatively high numbers of single-celled, north-seeking magne-totactic bacteria as expected. When we moved to the oppositeside of the drop to look for south-seeking cells, we notedorganisms that were identical in size, morphology, and style ofmotility to the MMP (see Video S1 in the supplemental ma-terial). Most of the described MMPs range from 3 to 12 �m indiameter and consist of about 10 to 40 individual Gram-neg-ative cells (5, 45, 46). The size range of the MMP-like aggre-gates from Pyramid Lake and GL1 to GL3 was 7 to 11 �m, andthe average was 7.5 � 1.0 �m (n � 24) (Fig. 1). Disruption ofthe MMPs into their component cells is known to occur whenthey are left under the microscope for periods ranging from 15

VOL. 76, 2010 NONMARINE NONMAGNETOTACTIC MULTICELLULAR PROKARYOTES 3221

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

min to 1 h or when they are placed in distilled water (45, 46).This allows for accurate counting of individual cells without theuse of electron microscopy. The nMMPs from Pyramid Lakeand the GBS pools did not disaggregate as easily as normalMMPs even when left in wet mounts or hanging drops for verylong periods of time. Thus, we were unable to obtain a largenumber of cell counts for individual nMMPs. In addition, whennMMPs did disrupt, individual cells remained tightly associ-ated with each other (Fig. 1A). For the nMMPs whose cells wecould count, there were between 40 and 60 cells per nMMP(Fig. 1A). It seems that the overall structure of the nMMPaggregates is more robust than that previously observed forMMPs, possibly due to the connection between cells or thematerial holding them together being stronger than in MMPsexamined thus far. Like other MMPs, an internal acellularcompartment (30, 33) appeared to be present in these nMMPsif the light microscope was focused appropriately (Fig. 1C).Transmission electron microscopy confirmed the multicellularnature of the nMMPs as well as the tight association of thecells (Fig. 2A and B) and the presence of flagella (Fig. 2B).

The nMMPs did not accumulate at the edge of the drop aswould typical MMPs and other magnetotactic bacteria in amagnetic field, nor did they exhibit the unusual “ping-pong”behavior (22, 45, 46) or “escape” motility (33), motility in amagnetic field described as spontaneous linear reversals ofhundreds of micrometers with decreasing speed, followed byforward swimming with increasing velocity (22). Generally thisoccurs when MMPs swim and reach the edge of the hangingdrop in the magnetic field. Instead, the nMMPs showed nomagnetotactic behavior and migrated away from the light tothe other side of the drop after several minutes (see Video S2in the supplemental material). When the light beam was placedon them again at the north side of the drop, they swam towardthe south side. This negative phototactic response to whitelight was observed for nMMPs collected from all samplingsites.

Using the hanging drop technique and colored filters, spe-cific wavelengths of light were tested for the ability to inducethe negative phototactic response in nMMPs. Exposure of adrop of water containing nMMPs to wavelengths of 435 and

FIG. 1. Differential interference contrast (DIC) light micrographs of the nonmagnetotactic multicellular prokaryotes (nMMPs). (A) Timecourse microscopy showing disaggregation of the nMMP. (B) Several nMMPs demonstrating variation in size of the organism. (C) The internalacellular compartment (arrows) of the nMMPs can be seen when the nMMPs are made immotile using fluorescence with a Lumar 49 filter set (thesame filter set used for cells stained with 4�,6-diamidino-2-phenylindole [DAPI]; emission BP445/50), and the microscope is focused appropriately.

FIG. 2. Transmission electron microscope (TEM) images of nMMPs and an MMP. (A) TEM image of an unstained nMMP showing theabsence of magnetosomes. (B) TEM image of a negatively stained nMMP showing the numerous flagella (at dark arrows) similar to that of theMMP shown in panel C. Again, note the absence of magnetosomes. (C) For a comparison, a TEM image of an MMP from the hypersaline SaltonSea (California) demonstrating the presence of greigite-containing magnetosomes (white arrows).

3222 LEFEVRE ET AL. APPL. ENVIRON. MICROBIOL.

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

480 nm for 5 min resulted in nMMP migration to the extremeedge of the other side of the drop (the “dark” side of the drop)(Fig. 3). Exposure to wavelengths of 530 and 640 nm undersimilar conditions showed little if any effect (Fig. 3).

Photoconcentration of nMMPs. The negative phototacticresponse in the nMMPs appeared robust enough to exploit fortheir concentration as well as to decrease or eliminate thenumber of contaminating bacteria. Because the nMMPs wereonly a very small component of the population of bacteria inthe samples, this would prove important for DNA extractionfor 16S rRNA gene cloning and sequencing and also for elec-tron microscopic examination. To test this possibility, we mod-ified the capillary racetrack of Wolfe et al. (60) that was orig-inally used for the concentration and/or physical isolation ofmagnetotactic bacteria to be used as inocula for cultivation(50), for DNA extraction (18, 56), and for electron microscopicexamination (36). We used the same general setup and appa-ratus of the capillary racetrack as follows (Fig. 4): a sterilePasteur pipette sealed at the thin end and containing a cottonplug was filled with filter-sterilized water from the sample up tothe cotton filter. Typically, 1 ml of mud and water from thesample was used to fill the open end of the pipette above the

cotton plug. In the original racetrack, magnets were placed atthe ends of and parallel to the racetrack to create a magneticfield to separate north- or south-seeking magnetotactic bacte-ria (60). In this case, magnets were not used and a strong lightsource, a Fiber-L-Lite high-intensity illuminator series 180(Dolan-Jenner Industries, Boxborough, MA), set at the maxi-mum intensity, was placed at the open end of the racetrack(Fig. 4). The sealed end of the capillary was covered with blacktape to negate any effects of extraneous light in the room. Theidea was that the motile nMMPs exposed to the strong lightwould be driven through the cotton plug to the dark sealed endof the pipette, where they would accumulate. Contaminatingcells would be significantly delayed in passage into the sterilewater by the cotton plug. All light racetracks were operated for30 min. Purified nMMPs were retrieved by breaking off thesealed end of the pipette and quantified as described abovewhere appropriate or used for DNA extraction for phyloge-netic analysis.

The system was tested to see if we could quantitatively provethat light could be used to concentrate the nMMPs. In thisexperiment, all racetracks were cut so the capillary end of thepipette was 8 cm in length (the distance from the cotton plugat the junction to the sealed end). The samples tested werethose from pool GL1 from GBS and Pyramid Lake. After theopen end containing 1 ml of the sample was exposed to thelight for 30 min, the dark sealed end of the capillary wasbroken off approximately 2 cm from the end, and the watercontaining nMMPs was removed and brought up to a finalvolume of 50 �l from which the hanging drops (3 per race-track) were made and nMMPs counted. Results from theseexperiments which clearly indicate that intense white light canbe effectively used to photoconcentrate nMMPs are shown inFig. 5. The relatively large standard deviations in the counts ofnMMPs in the racetracks in Fig. 5 reflect the variability innumbers of nMMPs in the original samples.

Testing for magnetotaxis and the presence of magneto-somes. Because magnetosomes in the MMP are often not asorganized as well as they are in other magnetotactic bacteria,

FIG. 3. Histogram showing the total number of nMMPs that swamto and accumulated at the dark far edge of a 5-�l drop exposed at theopposite end to a specific wavelength of light for 5 min. Data arerepresented as the means and standard deviations from 5 replicates.

FIG. 4. Light “racetrack” modified from the magnetic capillary racetrack of Wolfe et al. (60) that was originally used to concentratemagnetotactic bacteria. Here the light racetrack was used to photoconcentrate nMMPs. The Pasteur pipette is filled with water and sediments, andthe capillary is filled with filtered water; the light source leads the nMMPs through the cotton plug in the direction of the capillary, where theyaccumulate when they reach the darkness. Bar, 3 cm.

VOL. 76, 2010 NONMARINE NONMAGNETOTACTIC MULTICELLULAR PROKARYOTES 3223

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

the lack of magnetotaxis in nMMPs could be due to (i) thepresence of disorganized magnetosomes and/or magnetosomechains that could result in the lack of a north and south end ofthe organism (i.e., a magnetic dipole), (ii) too few magneto-somes to elicit a magnetotactic response, or (iii) a lack ofmagnetosomes altogether. Examination of nMMPs from allsites with electron microscopy showed the latter to be the case(Fig. 2). We were unable to confirm the presence of a singlemagnetosome in any nMMP.

Phylogeny of nMMPs. Three of the seven 16S rRNA genescloned and sequenced from photoconcentrated racetracks ofnMMPs from site GL1, two of the eight sequenced for siteGL3, and two of the eight sequenced for Pyramid Lake wereclosely related (�99.3% sequence similarity) to sequences ofthe MMP group in the Deltaproteobacteria class (Fig. 6). Be-cause the sequence divergence between the different nMMPsclones was very small, we considered that it may be the resultof errors in sequencing. The minimum difference in 16S rRNAgene sequence between nMMPs was 3/1,527 bases and themaximum was 11/1,527 bases. The maximum possible error weobtained with our sequencing facility was 5 bases on an entire16S rRNA gene sequence of an axenic culture. Thus, it ispossible that errors in sequencing might contribute to the ob-served sequence divergence in the nMMPs in our study. How-ever, even taking this possibility into consideration, the se-quences from the nMMPs clearly form a new phylogeneticbranch within the MMP group (Fig. 6) and show that thenMMPs sequenced in our study probably represent a singlespecies within the group.

DISCUSSION

The uncultured microorganism known as the MMP has in-trigued microbiologists for a number of years because of theimplications related to its obligately multicellular nature andthe fact that it was the first magnetotactic bacterium discoveredthat biomineralized greigite (39) from nonmagnetic iron sul-

fide precursors (43, 44) and that could incorporate a metalother than iron (copper) in its magnetosomes (6). More spe-cifically, the MMP’s multicellularity suggests to many thatsome kind of intercellular communication must occur in thisorganism for it to function—for example, the apparent coor-dination of the numerous flagella for directed motility—and todivide.

The first important question to be answered is whethernMMPs simply share the same apparently unusual morphologyand ultrastructure of MMPs or are phylogenetically related tothem; i.e., are they actually an unrecognized form of MMP?Phylogenetic analyses of the nMMPs clearly demonstrate thatthey are members of the Deltaproteobacteria as part of a largegroup of sulfate-reducing bacteria and that they are closelyrelated to MMPs. However, the nMMPs appear to form aseparate, coherent phylogenetic branch within the MMPgroup. Considering the degree of 16S rRNA gene sequencesimilarity (�99.3%) between the nMMPs and the sequencedifference between these and the MMPs (�95.8%), nMMPsappear to represent a new form of MMP that consists of asingle species that is distinct from the several species repre-sented by the MMPs (53).

MMP-like organisms had previously only been found in orassociated with marine environments with salinities that arerelatively high even compared to hypersaline environments (2,30, 32, 40). For this reason, many consider the MMP to be anexclusively marine organism (53). Our results show that MMP-like organisms are not confined to marine environments, butbecause the aquatic habitats that we studied are still saline (i.e.,brackish), that does not eliminate the possibility that MMP-like organisms require elevated levels of certain salts comparedto freshwater bacteria. Different types of MMPs appear tohave distinct salinity ranges for growth and survival that wouldpresumably reflect the extent of any salt requirement. Forexample, viability studies involving the MMP “CandidatusMagnetoglobus multicellularis,” from the hypersaline Ara-ruama lagoon in Brazil (2), showed that they died when ex-posed to salinities of �55 and �40 ppt in microcosms (40).Interestingly, however, viable organisms were found in thelagoon even when the salinity was 60 ppt, although in lowernumbers when the salinity was intermediate between 40 and 55ppt (average salinity of the lagoon, 40 ppt). Nonetheless, theknown salinities in which MMP-like organisms exist have nowbeen extended to as low as 5 ppt. Pure cultures will be requiredto determine whether specific salts at specific concentrationsare required for growth and survival of the MMP-like organ-isms.

The nMMPs did not exhibit magnetotaxis, nor did they bio-mineralize magnetosomes of any type. Environmental condi-tions are also known to affect magnetosome production, and itis possible that chemical and redox conditions present at thesampling sites are not conducive to intracellular greigite bio-mineralization in magnetosomes. This notion is unlikely be-cause other greigite-producing magnetotactic bacteria, as wellas magnetite producers, were also present in the same samples.The greigite producers in these samples, large rod-shaped cellswith a double chain of magnetosomes, were similar to thoseoften found together with MMPs from marine sites (5, 6; seealso Video S2 in the supplemental material).

Some magnetotactic bacteria are known to frequently lose

FIG. 5. Effect of light on nMMPs in capillary light racetracks(shown in Fig. 4). Histogram of numbers of nMMPs refers to nMMPsfrom samples from Great Boiling Springs pool GL1 (GL1) and Pyra-mid Lake (PL) that swam to and accumulated at the end of the lightracetrack after 30 min of exposure to light or darkness (negativecontrol). Data are presented as the means and standard deviations oftriplicate counts of nMMPs from triplicate racetracks.

3224 LEFEVRE ET AL. APPL. ENVIRON. MICROBIOL.

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

the phenotypic trait of magnetotaxis and the ability to biomin-eralize magnetosomes in culture (14, 20, 34, 58). It has beenshown that the genes encoding magnetosome (membrane)proteins are localized on a 130-kb magnetosome genomic is-land (MAI) in Magnetospirillum gryphiswaldense (48, 58). Sim-ilar magnetosome islands appear to be present in the genomesof other Magnetospirillum strains (20, 27, 28), as well as in themarine vibrio strain MV-1 (28). The MAIs in these organismsare flanked and interrupted in some cases by several types ofmobile genetic elements, including direct repeats, insertionsequences, integrases, transposases, proximal transfer RNAs,and areas of atypical G�C content, which are responsible forthe mobilization of the island (28). Because of these mobileelements, the MAI, like other genomic islands, is likely capableof horizontal gene transfer (13, 23). In addition, genomic is-lands not only have a strong tendency to be deleted fromgenomes with high frequency, they also can undergo duplica-

tions, amplifications, and rearrangements (13, 23, 25). Hori-zontal gene transfer of the MAI may explain why the trait ofmagnetotaxis is widespread, occurring in many diverse, unre-lated bacteria, and why the trait is easily lost in various strainsof magnetotactic bacteria. Strong evidence from tetranucle-otide usage patterns in Magnetospirillum species and strainMV-1 showing that the MAIs in these organisms are trans-ferred by horizontal gene transfer was recently described (28).Another suggestive piece of evidence for horizontal gene trans-fer of the MAI is the discovery of Magnetospirillum species thatare not magnetotactic and do not biomineralize magnetosomes(10, 21, 51). Amann et al. (3) point out that the 98% 16S rRNAgene similarity between nonmagnetotactic Aquaspirillum poly-morphum and M. gryphiswaldense is significantly greater thanthat between M. magnetotacticum and M. gryphiswaldense, andthey pose the question of whether A. polymorphum representsa Magnetospirillum strain that recently lost the genes required

FIG. 6. Phylogenetic tree based on 16S rRNA gene sequences showing the phylogenetic position of the nMMPs in the Deltaproteobacteria class.GL1 and GL3 refer to two different pools from the Great Boiling Springs site, and PL refers to Pyramid Lake where the nMMPs were collected.Bootstrap values at nodes are percentages of 1,000 replicates. The magnetotactic bacterium Desulfovibrio magneticus, Desulfonema limicola, andDesulfosarcina variabilis (all from other groups in the Deltaproteobacteria class) were used to root the tree. GenBank accession numbers are shownin parentheses. Bar, 2% sequence divergence.

VOL. 76, 2010 NONMARINE NONMAGNETOTACTIC MULTICELLULAR PROKARYOTES 3225

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

for magnetotaxis, or whether M. gryphiswaldense recently ac-quired these genes. A similar situation seems to exist herebetween the MMPs and nMMPs. The nMMPs clearly repre-sent a different species of the MMPs that may have lost thegenes for greigite biomineralization or has not (yet?) acquiredthem. If this is true, then the genes for greigite biomineraliza-tion might be organized similarly (i.e., as an MAI) to those formagnetite production in the magnetospirilla.

One of the most interesting behaviors we observed in thenMMPs was their strong negative phototactic reaction to whitelight. This response was robust enough to exploit for the pho-toconcentration of nMMPs for DNA extraction for phyloge-netic analysis and for the preparation of grids for electronmicroscopy. When nMMPs were tested for this negative pho-totactic response to specific wavelengths of visible light, onlywavelengths of �480 nm (blue to violet) elicited the behavior.Longer wavelengths (�530 nm; green to red) did not. There isphylogenetic, genetic, and ecological evidence that stronglysuggests that all forms of MMPs are anaerobic sulfate-reducingbacteria (2, 5, 12, 53, 59). The most obvious explanation forthis negative phototactic response is that light would drivenMMPs, and perhaps MMPs, in the environment toward andinto the anaerobic zone of aquatic habitats where they arefound in their highest numbers (54). It is significant that theshorter wavelengths of visible light that caused the response,�480 nm (blue to violet), are those that generally penetratethe water column the deepest (9). The negative phototacticresponse of the nMMPs is clearly efficient, and it is interestingthat in this case, it might function similarly to magnetotaxis.That is, if light causes the nMMPs in nature to swim more orless vertically, then, like magnetotaxis (19), it would help toreduce a three-dimensional search problem to a one-dimen-sional search problem for an organism that must locate andmaintain a position where specific chemical and redox condi-tions are optimal in a vertical chemical and redox gradientcommon in aquatic habitats. In this way, negative phototaxis inthis case might increase the efficiency of chemotaxis as doespolar magneto-aerotaxis (19). Although phototaxis as a trait isdistributed widely among prokaryotes (26), as far as we know,this is the first report of a form of phototaxis in the Deltapro-teobacteria class.

Other prokaryotes show a similar response to blue light forapparently different reasons. For example, the halophilic pur-ple bacterium Ectothiorhodospira (Halorhodospira) halophilaexhibits a negative phototactic response to blue light, presum-ably to avoid the damaging effects of UV light (8, 55). Thephotoactive yellow protein PYP (41, 42) is thought to be thephotoreceptor responsible for this negative photoresponse,and it appears that many prokaryotes synthesize similar pro-teins or other blue light receptors (8). Cells of Escherichia colidisplay tumbling behavior upon short exposure (1 s) to intenseblue light (8). After tumbling, cells run and soon becomeimmotile and die, presumably from the formation of reactiveoxygen species (8).

Despite the fact that no MMP-like organisms have yet beenisolated and cultivated, a great deal of information regardingthese unusual prokaryotes has been obtained through the useof culture-independent means and ecological studies. Resultspresented in this paper significantly advance and extend whatwe know about these microorganisms. Evidence in this paper

clearly shows that the MMP group contains forms that are notmagnetotactic, the nMMPs, which represent a new subgroup ofthe MMP group, probably a single species. Our results alsoshow that MMP-like organisms are not restricted to marineenvironments, although, because the habitats we describe arestill brackish by definition, elevated levels of certain salts maybe required for these organisms to exist. Nonetheless, thelower limit of salinity for these types of organisms to exist isnow �5 ppt, which now gives cause to look for them in fresh-water environments. The discovery of the nMMPs also raisesmany important questions. MMPs also do not evidently needto be magnetotactic and biomineralize magnetosomes, whichleads to the question of the function of these organelles in theMMP. It seems possible that the nMMP’s negative phototacticresponse might replace magnetotaxis in increasing the effi-ciency of chemotaxis in these organisms. It would be interest-ing now to know whether MMPs show this same phototacticresponse. It seems apparent that MMP-like microorganismswill continue to intrigue microbiologists in the future.

ACKNOWLEDGMENTS

We dedicate this paper to the memory of the late Karl Canter, aphysicist whose research focus took on a major microbiological com-ponent when he was introduced to the MMP.

We thank David and Sandie Jamieson for access to the GBS fieldsite and J. Dodsworth for help collecting samples.

This work was supported by U.S. National Science Foundation grantEAR-0920718 to D.A.B. U.L. and F.A. acknowledge partial financialsupport from the National Council for Scientific and TechnologicalDevelopment (CNPq) of Brazil.

REFERENCES

1. Abreu, F., K. T. Silva, J. L. Martins, and U. Lins. 2006. Cell viability inmulticellular magnetotactic prokaryotes. Int. Microbiol. 9:267–271.

2. Abreu, F., J. L. Martins, T. S. Silveira, C. N. Keim, H. G. P. Lins de Barros,F. J. G. Filho, and U. Lins. 2007. ‘Candidatus Magnetoglobus multicellu-laris,’ a multicellular, magnetotactic prokaryote from a hypersaline environ-ment. Int. J. Syst. Evol. Microbiol. 57:1318–1322.

3. Amann, R., J. Peplies, and D. Schuler. 2006. Diversity and taxonomy ofmagnetotactic bacteria, p. 25–36. In D. Schuler (ed.), Magnetoreception andmagnetosomes in bacteria, vol. 3. Springer, Berlin, Germany.

4. Anderson, J. P. 1978. A geochemical study of the southwest part of the BlackRock Desert and its geothermal areas; Washoe, Pershing, and HumboldtCounties, Nevada. Colo. School Mines Q. 73:15–22.

5. Bazylinski, D. A., R. B. Frankel, A. J. Garrat-Reed, and S. Mann. 1990.Biomineralisation of iron sulfides in magnetotactic bacteria from sulfidicenvironments, p. 239–255. In R. B. Frankel and R. P. Blakemore (ed.), Ironbiominerals. Plenum Press, New York, NY.

6. Bazylinski, D. A., A. J. Garrat-Reed, A. Abedi, and R. B. Frankel. 1993.Copper association with iron sulfide magnetosomes in a magnetotactic bac-terium. Arch. Microbiol. 160:35–42.

7. Benson, L. V., D. R. Currey, R. I. Dorn, K. R. Lajoie, C. G. Oviatt, S. W.Robinson, G. I. Smith, and S. Stine. 1990. Chronology of expansion andcontraction of four Great Basin lake systems during the past 35,000 years.Palaeogeogr. Palaeocl. 78:241–286.

8. Braatsch, S., and G. Klug. 2004. Blue light perception in bacteria. Photo-synth. Res. 79:45–57.

9. Braatsch, S., O. V. Moskvin, G. Klug, and M. Gomelsky. 2004. Responses ofthe Rhodobacter sphaeroides transcriptome to blue light under semiaerobicconditions. J. Bacteriol. 186:7726–7735.

10. Coates, J. D., U. Michaelidou, R. A. Bruce, S. M. O’Connor, J. N. Crespi, andL. A. Achenbach. 1999. Ubiquity and diversity of dissimilatory (per)chlorate-reducing bacteria. Appl. Environ. Microbiol. 65:5234–5241.

11. Costa, K. C., J. B. Navarro, E. L. Shock, C. L. Zhang, D. Soukup, and B. P.Hedlund. 2009. Microbiology and geochemistry of great boiling and mud hotsprings in the United States Great Basin. Extremophiles 13:447–459.

12. DeLong, E. F., R. B. Frankel, and D. A. Bazylinski. 1993. Multiple evolu-tionary origins of magnetotaxis in bacteria. Science 259:803–806.

13. Dobrindt, U., B. Hochhut, U. Hentschel, and J. Hacker. 2004. Genomicislands in pathogenic and environmental microorganisms. Nat. Rev. Micro-biol. 2:414–424.

14. Dubbels, B. L., A. A. DiSpirito, J. D. Morton, J. D. Semrau, J. N. Neto, and

3226 LEFEVRE ET AL. APPL. ENVIRON. MICROBIOL.

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

D. A. Bazylinski. 2004. Evidence for a copper-dependent iron transportsystem in the marine, magnetotactic bacterium strain MV-1. Microbiology150:2931–2945.

15. Esquivel, D. M. S., H. G. P. Lins de Barros, M. Farina, P. H. A. Aragao, andJ. Danon. 1983. Microorganismes magnetotactiques de la region de Rio deJaneiro. Biol. Cell 47:227–234.

16. Farina, M., H. Lins de Barros, D. Motta de Esquivel, and J. Danon. 1983.Ultrastructure of a magnetotactic microorganism. Biol. Cell 48:85–88.

17. Farina, M., D. Motta de Esquivel, and H. G. P. Lins de Barros. 1990.Magnetic iron-sulphur crystals from a magnetotactic microorganism. Nature343:256–258.

18. Flies, C. B., H. M. Jonkers, D. de Beer, K. Bosselmann, M. E. Bottcher, andD. Schuler. 2005. Diversity and vertical distribution of magnetotactic bacte-ria along chemical gradients in freshwater microcosms. FEMS Microbiol.Ecol. 52:185–195.

19. Frankel, R. B., D. A. Bazylinski, M. S. Johnson, and B. L. Taylor. 1997.Magneto-aerotaxis in marine, coccoid bacteria. Biophys. J. 73:994–1000.

20. Fukuda, Y., Y. Okamura, H. Takeyama, and T. Matsunaga. 2006. Dynamicanalysis of a genomic island in Magnetospirillum sp. strain AMB-1 revealshow magnetosome synthesis developed. FEBS Lett. 580:801–812.

21. Geelhoed, J. S., D. Y. Sorokin, E. Epping, T. P. Tourova, H. L. Banciu, G.Muyzer, A. J. M. Stams, and M. C. M. van Loosdrecht. 2009. Microbialsulfide oxidation in the oxic-anoxic transition zone of freshwater sediment:involvement of lithoautotrophic Magnetospirillum strain J10. FEMS Micro-biol. Ecol. 70:54–65.

22. Greenberg, M., K. Canter, I. Mahler, and A. Tornheim. 2005. Observation ofmagnetoreceptive behavior in a multicellular magnetotactic prokaryote inhigher than geomagnetic fields. Biophys. J. 88:1496–1499.

23. Hacker, J., and J. B. Kaper. 2000. Pathogenicity islands and the evolution ofmicrobes. Annu. Rev. Microbiol. 54:641–679.

24. Hall, T. A. 1999. BioEdit: a user-friendly biological sequence alignmenteditor and analysis program for Windows 95/98/NY. Nucl. Acids Symp. Ser.41:95–98.

25. Hsiao, W., I. Wan, S. J. Jones, and F. S. L. Brinkman. 2003. IslandPath:aiding detection of genomic islands in prokaryotes. Bioinformatics 19:418–420.

26. Jekely, G. 2009. Evolution of phototaxis. Philos. Trans. R. Soc. Lond. B364:2795–2808.

27. Jogler, C., and D. Schuler. 2007. Genetic analysis of magnetosome biomin-eralization, p. 133–161. In D. Schuler (ed.), Magnetoreception and magne-tosomes in bacteria, vol. 3. Springer, Berlin, Germany.

28. Jogler, C., M. Kube, S. Schubbe, S. Ullrich, H. Teeling, D. A. Bazylinski, R.Reinhardt, and D. Schuler. 2009. Comparative analysis of magnetosomegene clusters in magnetotactic bacteria provides further evidence for hori-zontal gene transfer. Environ. Microbiol. 11:1267–1277.

29. Jones, J. C. 1914. The geologic history of Lake Lahontan. Science 40:827–830.

30. Keim, C. N., F. Abreu, U. Lins, H. G. P. Lins de Barros, and M. Farina. 2004.Cell organization and ultrastructure of a magnetotactic multicellular organ-ism. J. Struct. Biol. 145:254–262.

31. Keim, C. N., U. Lins, and M. Farina. 2003. Iron oxide and iron sulfidecrystals in magnetotactic multicellular aggregates. Acta Microsc. 12:3–4.

32. Keim, C. N., J. L. Martins, F. Abreu, A. S. Rosado, H. G. P. Lins de Barros,R. Borojevic, U. Lins, and M. Farina. 2004. Multicellular life cycle of mag-netotactic multicellular prokaryotes. FEMS Microbiol. Lett. 240:203–208.

33. Keim, C. N., J. L. Martins, H. G. P. Lins de Barros, U. Lins, and M. Farina.2007. Structure, behaviour, ecology and diversity of multicellular magneto-tactic prokaryotes, p. 103–132. In D. Schuler (ed.), Magnetoreception andmagnetosomes in bacteria, vol. 3. Springer, Berlin, Germany.

34. Komeili, A., Z. Li, D. K. Newman, G. J. Jensen. 2006. Magnetosomes are cellmembrane invaginations organized by the actin-like protein MamK. Science311:242–245.

35. Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115–175. In E. Stackebrandt,and M. Goodfellow (ed.), Nucleic acid techniques in bacterial systematics.Wiley & Sons, Chichester, United Kingdom.

36. Lefevre, C., A. Bernadac, N. Pradel, L. Wu, K. Yu-Zhang, T. Xiao, J.-P.Yonnet, A. Lebouc, T. Song, and Y. Fukumori. 2007. Characterization ofMediterranean magnetotactic bacteria. J. Ocean Univ. China 6:355–359.

37. Lins, U., and M. Farina. 1999. Organisation of cells in magnetotactic mul-ticellular aggregates. Microbiol. Res. 154:9–13.

38. Lins, U., C. N. Keim, F. F. Evans, P. R. Buseck, and M. Farina. 2007.Magnetite (Fe3O4) and greigite (Fe3S4) crystals in multicellular magneto-tactic prokaryotes. Geomicrobiol. J. 24:43–50.

39. Mann, S., N. H. C. Sparks, R. B. Frankel, D. A. Bazylinski, and H. W.

Jannasch. 1990. Biomineralization of ferrimagnetic greigite (Fe3S4) and ironpyrite (FeS2) in a magnetotactic bacterium. Nature 343:258–261.

40. Martins, J. L., T. S. Silveira, K. T. Silva, and U. Lins. 2009. Salinity depen-dence of the distribution of multicellular magnetotactic prokaryotes in ahypersaline lagoon. Int. Microbiol. 12:193–201.

41. Meyer, T. E. 1985. Isolation and characterization of soluble cytochromes,ferredoxins and other chromophoric proteins from the halophilic phototro-phic bacterium Ectothiorhodospira halophila. Biochim. Biophys. Acta 806:175–183.

42. Meyer, T. E., J. C. Fitch, R. G. Bartsch, G. Tollin, and M. A. Cusanovich.1990. Soluble cytochromes and a photoactive yellow protein from the mod-erately halophilic purple phototrophic bacterium, Rhodospirillum salexigens.Biochim. Biophys. Acta 1016:364–370.

43. Posfai, M., P. R. Buseck, D. A. Bazylinski, and R. B. Frankel. 1998. Reactionsequence of iron sulfide minerals in bacteria and their use as biomarkers.Science 280:880–883.

44. Posfai, M., P. R. Buseck, D. A. Bazylinski, and R. B. Frankel. 1998. Ironsulfides from magnetotactic bacteria: structure, composition and phase tran-sitions. Am. Mineral. 83:1469–1481.

45. Rodgers, F. G., R. P. Blakemore, N. A. Blakemore, R. B. Frankel, D. A.Bazylinski, D. Maratea, and C. Rodgers. 1990. Intercellular structure in amany-celled magnetotactic prokaryote. Arch. Microbiol. 145:18–22.

46. Rodgers, F. G., R. P. Blakemore, N. A. Blakemore, R. B. Frankel, D. A.Bazylinski, D. Maratea, and C. Rodgers. 1990. Intercellular junctions, mo-tility and magnetosome structure in a multicellular magnetotactic pro-karyote, p. 231–237. In R. B. Frankel and R. P. Blakemore (ed.), Ironbiominerals. Plenum Press, New York, NY.

47. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new methodfor reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–425.

48. Schubbe, S., M. Kube, A. Scheffel, C. Wawer, U. Heyen, A. Meyerdierks,M. H. Madkour, F. Mayer, R. Reinhardt, and D. Schuler. 2003. Character-ization of a spontaneous nonmagnetic mutant of Magnetospirillum gryphi-swaldense reveals a large deletion comprising a putative magnetosome island.J. Bacteriol. 185:5779–5790.

49. Schuler, D. 2002. The biomineralization of magnetosomes in Magnetospiril-lum gryphiswaldense. Int. Microbiol. 5:209–214.

50. Schuler, D., S. Spring, and D. A. Bazylinski. 1999. Improved technique forthe isolation of magnetotactic spirilla from a freshwater sediment and theirphylogenetic characterization. Syst. Appl. Microbiol. 22:466–471.

51. Shinoda, Y., Y. Sakai, M. Ue, A. Hiraishi, and N. Kato. 2000. Isolation andcharacterization of a new denitrifying spirillum capable of anaerobic degra-dation of phenol. Appl. Environ. Microbiol. 66:1286–1291.

52. Silva, K. T., F. Abreu, F. P. Almeida, C. N. Keim, M. Farina, and U. Lins.2007. Flagellar apparatus of south seeking many celled magnetotactic pro-karyotes. Microsc. Res. Tech. 70:10–17.

53. Simmons, S. L., and K. J. Edwards. 2007. Unexpected diversity in popula-tions of the many-celled magnetotactic prokaryote. Environ. Microbiol.9:206–215.

54. Simmons, S. L., S. M. Sievert, R. B. Frankel, D. A. Bazylinski, and K. J.Edwards. 2004. Spatiotemporal distribution of marine magnetotactic bacte-ria in a seasonally stratified coastal salt pond. Appl. Environ. Microbiol.70:6230–6239.

55. Sprenger, W. W., W. D. Hoff, J. P. Armitage, and K. J. Hellingwerf. 1993. Theeubacterium Ectothiorhodospira halophila is negatively phototactic, with awavelength dependence that fits the absorption spectrum of the photoactiveyellow protein. J. Bacteriol. 175:3096–3104.

56. Spring, S., R. Amann, W. Ludwig, K.-H. Schleifer, and N. Petersen. 1992.Phylogenetic diversity and identification of nonculturable magnetotactic bac-teria. Syst. Appl. Microbiol. 15:116–122.

57. Tamura, K., J. Dudley, M. Nei, and S. Kumar. 2007. MEGA4: MolecularEvolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol.Evol. 24:1596–1599.

58. Ullrich, S., M. Kube, S. Schubbe, R. Reinhardt, and D. Schuler. 2005 Ahypervariable 130-kilobase genomic region of Magnetospirillum gryphiswal-dense comprises a magnetosome island which undergoes frequent rearrange-ments during stationary growth. J. Bacteriol. 187:7176–7184.

59. Wenter, R., G. Wanner, D. Schuler, and J. Overmann. 2009. Ultrastructure,tactic behaviour and potential for sulfate reduction of a novel multicellularmagnetotactic prokaryote from North Sea sediments. Environ. Microbiol.11:1493–1505.

60. Wolfe, R. S., R. K. Thauer, and N. Pfennig. 1987. A “capillary racetrack”method for isolation of magnetotactic bacteria. FEMS Microbiol. Ecol. 45:31–35.

VOL. 76, 2010 NONMARINE NONMAGNETOTACTIC MULTICELLULAR PROKARYOTES 3227

on February 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from