1

Probing the nanostructure and arrangement of bacterial 1

magnetosomes by small-angle x-ray scattering 2

3

Sabine Rosenfeldt# 1,2, Cornelius N. Riese# 3, Frank Mickoleit3, Dirk Schüler3, Anna S. 4

Schenk*1,4 5

6

# These authors contributed equally to this work 7

8

1Bavarian Polymer Institute (BPI), 2Physical Chemistry 1, 3Dept. Microbiology, 9

4Physical Chemistry - Colloidal Systems; University of Bayreuth, D-95447 Bayreuth 10

11

* corresponding author: Jun.-Prof. Dr. Anna S. Schenk, Physical Chemistry - 12

Colloidal Systems, University of Bayreuth, D-95447 Bayreuth; E-Mail: 13

anna.schenk@uni-bayreuth.de; Tel.: +49-921-55-3915 14

15

16

Abstract 17

Magnetosomes are membrane-enveloped single-domain ferromagnetic nanoparticles 18

enabling the navigation of magnetotactic bacteria along magnetic field lines. Strict 19

control over each step of biomineralization generates particles of high crystallinity, 20

strong magnetization and remarkable uniformity in size and shape, particularly 21

interesting for many biomedical and biotechnological applications. However, to 22

understand the physicochemical processes involved in magnetite biomineralization, 23

precise and permanent monitoring of particle production is required. Common 24

techniques such as transmission electron microscopy (TEM) or Fe measurements 25

AEM Accepted Manuscript Posted Online 11 October 2019Appl. Environ. Microbiol. doi:10.1128/AEM.01513-19Copyright © 2019 Rosenfeldt et al.This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

2

only allow for semi-quantitative assessment of the magnetosome formation without 26

routinely revealing quantitative structural information. In this study, lab-based small-27

angle x-ray scattering (SAXS) is explored as a means to monitor the different stages 28

of magnetosome biogenesis in the model organism Magnetospirillum 29

gryphiswaldense. SAXS is evaluated as a quantitative stand-alone technique to 30

analyse the size, shape and arrangement of magnetosomes in cells cultivated under 31

different growth conditions. By applying a simple and robust fitting procedure based 32

on linearly aligned spheres, it is demonstrated that the SAXS datasets contain 33

information on both the diameter of the inorganic crystal and the protein-rich 34

magnetosome membrane. The analyses reveal a narrow particle size distribution with 35

an overall magnetosome radius of 19 nm in Magnetospirillum gryphiswaldense. 36

Furthermore, the averaged distance between individual magnetosomes is 37

determined, revealing a chain-like particle arrangement with a center-to-center 38

distance of 53 nm. Overall, these data demonstrate that SAXS can be used as a 39

novel stand-alone technique allowing for the at-line monitoring of magnetosome 40

biosynthesis, thereby providing accurate information on the particle nanostructure. 41

42

43

Importance 44

This study explores lab-based small-angle x-ray scattering (SAXS) as a novel 45

quantitative stand-alone technique to monitor the size, shape and arrangement of 46

magnetosomes during different stages of particle biogenesis in the model organism 47

Magnetospirillum gryphiswaldense. The SAXS datasets contain volume-averaged, 48

statistically accurate information on both the diameter of the inorganic nanocrystal 49

and the enveloping protein-rich magnetosome membrane. As a robust and non-50

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

3

destructive in-situ technique, SAXS can provide new insights into the 51

physicochemical steps involved in the biosynthesis of magnetosome nanoparticles as 52

well as their assembly into well-ordered chains. The proposed fit model can be easily 53

adapted to account for different particle shapes and arrangements produced by other 54

strains of magnetotactic bacteria, thus rendering SAXS a highly versatile method. 55

56

Keywords: Small-Angle X-Ray Scattering; SAXS; Magnetospirillum 57

gryphiswaldense; Magnetosomes; Magnetotactic bacteria, Magnetic nanostructure 58

59

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

4

Introduction 60

61

In order to facilitate their orientation and migration along the Earth’s magnetic field, 62

magnetotactic bacteria (MTB) biomineralize magnetosomes enabling them to 63

navigate towards favorable habitats in the oxic-anoxic transition zone. In the 64

alphaproteobacterium Magnetospirillum gryphiswaldense (Fig. 1a) these magnetic 65

nanoparticles consist of a cuboctahedral, monocrystalline core of chemically pure 66

magnetite (Fe3O4), enveloped by a protein-rich biological membrane (magnetosome 67

vesicle) aligned in a linear chain of about 40 particles (Fig. 1b) (1–4). Due to strict 68

control of each step of magnetosome biogenesis, particles with unprecedented 69

characteristics (with respect to crystallinity, particle size, shape and magnetization; 5, 70

6) are synthesized. However, despite of considerable efforts, the physicochemical 71

processes and steps involved in the biomineralization of well-ordered chains of 72

magnetite nanocrystals are still not understood. This is partially due to a lack of 73

practical lab-scale methods to assess bulk magnetosome formation in growing cells. 74

In addition, magnetosomes are of increasing interest as biogenic magnetic 75

nanoparticles as they have been already successfully explored in a wide range of 76

potential biotechnical and biomedical applications including magnetic hyperthermia 77

(7, 8), as contrast agents (9, 10) or particle based immune-assays (11–13). Yet, 78

bioproduction of magnetosome particles depends on highly controlled growth 79

conditions (14, 15) and requires precise and permanent monitoring of bulk 80

magnetosomes biomineralization during bacterial cultivation. 81

Commonly used proxies to estimate magnetosome formation, such as the magnetic 82

orientation of cells (cmag [16, 17]) and their iron content, lack sufficient specificity 83

and do not provide information about the size, structure and arrangement of particles. 84

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

5

In contrast, transmission electron microscopy (TEM) allows the precise determination 85

of particle numbers, sizes, and shapes as well as their chain assemblies. 86

Nonetheless, gathering statistically significant datasets from time-resolved 87

experiments on cultures is very laborious and time consuming. Furthermore, the 88

analysis is often hindered by preparation or drying artefacts. X-ray radiation based 89

techniques such as extended X-ray absorption fine structure (EXAFS) and X-ray 90

absorption near edge structure (XANES) were already applied for characterization of 91

magnetite biomineralization, however, quantitative TEM measurements were still 92

necessary to address magnetosome size distribution (18, 19). 93

As an alternative, small-angle X-ray scattering (SAXS) has been used to assess the 94

structural properties of bacterial magnetosomes (20–22). This technique represents a 95

volume-averaging method to study nanoscale interfaces based on electron density 96

heterogeneities within a sample. SAXS experiments performed on a laboratory based 97

setup provide access to size, shape and orientation of nanoscale objects spanning a 98

size range of 1–300 nm, and thus are ideally suited to study structural parameters 99

and aggregation behaviour of intermediates and final products formed during 100

magnetosome biomineralization, as the mature nanoparticles exhibit an average size 101

of about 40 nm (4, 12) and a substantial electron density contrast. During 102

magnetosome biogenesis, the forming particles characteristically undergo a 103

morphological transition leading to changes in the electron density distribution at 104

different growth stages. Such variations in contrast, e.g. evoked by vesicles 105

successively filling up with iron-rich magnetite mineral, are accessible by SAXS and 106

contribute to the scattering pattern in a characteristic way. 107

Krueger et al. (20) first combined small-angle neutron scattering (SANS) and SAXS 108

to investigate structural parameters of magnetite nanocrystals in magnetotactic 109

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

6

bacteria. Due to the restricted data quality of the individual scattering patterns, the 110

results were analysed by comparing the scattering profiles of samples with low and 111

high iron loading in the presence and absence of external magnetic fields. The 112

diameters of the magnetosomes biosynthesized during cultivation were estimated to 113

be 40 nm (20). Hoell et al. (21) later investigated M. gryphiswaldense by specialized 114

techniques based on both SANS and SAXS to access the size of the magnetosome 115

membrane and the magnetite crystals as well as their magnetization. More recently, 116

Molcan and co-workers (23) performed a comprehensive structural characterization 117

of magnetosome suspensions from M. magnetotacticum for potential hyperthermia 118

applications by applying a wide range of contrasting analytical techniques including 119

SAXS. The scattering data in the regime of low values of the scattering vector q were 120

fitted by a q-x power law, where the increase in the exponent x was interpreted as a 121

progressing fractal-like aggregation of magnetosomes. Orue et al. (22) focused on 122

the magnetic nanoarchitecture of magnetosomes from M. gryphiswaldense, but also 123

performed SAXS measurements on highly concentrated bacterial colloids. The 124

authors analysed the data sets by an indirect Fourier transformation method, which 125

usually relies on additional information regarding the geometry and size of possible 126

aggregate structures. Applying this strategy, Orue and coworkers (22) correlated the 127

first maximum of the pair distance distribution function to the size of one 128

magnetosome, and the subsequent maxima to the inter-particle distances of adjacent 129

particles. Although comparison with the results from direct visualization by cryo-130

electron tomography revealed a good agreement of the estimated magnetosome 131

size, a 15% lower value was estimated for the correlation distance. However, in all 132

SAXS studies on bacterial magnetosome, the method was used only on selected 133

specimens as a complementary tool to validate the results obtained from other 134

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

7

techniques such as TEM, electron tomography or SANS rather than as a stand-alone 135

analytical method. 136

In this study, SAXS is explored as a quantitative stand-alone bulk measurement 137

technique allowing time–efficient particle size analysis (in contrast to well-138

established, but laborious methods such as TEM). Despite the detailed structural 139

resolution of magnetosomes, which can be achieved by highly brilliant X-ray beams 140

produced at synchrotron sources (21), limited access as well as usually long 141

distances between production and measurement site render synchrotron SAXS 142

analysis unpractical for fast at-line magnetosome characterization. Accordingly, the 143

present study investigated concentrated cell suspensions of M. gryphiswaldense at 144

different growth stages using a lab-scale SAXS system with a rotating copper anode 145

as X-ray source. This in-house setup offers the advantage of easier access (in 146

contrast to synchrotron measurements) and close proximity to the fermentation 147

facility, thus enabling sampling and process control during magnetosome 148

biosynthesis. 149

150

151

Materials and Methods 152

153

Cultivation of M. gryphiswaldense 154

M. gryphiswaldense wild type was grown either aerobically or anaerobically at 28°C 155

in a BioflowTM 320 Eppendorf Bioprocess fermentor system (Jülich, Germany) using 156

a modified large-scale medium (14). Cells were harvested by low-spin centrifugation 157

(7,000 rpm, 10 min, 4°C; Beckmann Coulter Avanti J-26SXP, USA). Cell pellets were 158

washed and resuspended in 10 mM 2-[4-(2-hydroxyethyl)piperazin-1-159

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

8

yl]ethanesulfonic acid (HEPES), pH 7.0, before transferring the suspension into glass 160

capillaries ( = 1 mm, Hilgenberg, Germany) for further analysis. 161

162

Structural characterization of magnetosome nanoparticles 163

Transmission electron microscopy (TEM). 164

For TEM analysis of whole cells, specimens were directly deposited onto carbon-165

coated copper grids. TEM imaging was performed on a Jeol Jem 1400+ (Freising, 166

Germany) operated at an acceleration voltage of 80 kV. Images were recorded with a 167

Gatan Erlangshen ES500W CCD camera. Average particle sizes were measured by 168

data processing using the ImageJ software package v1.52i. 169

170

Small-Angle-X-ray Scattering (SAXS). 171

Nanostructure analysis of bacterial magnetososmes was performed on highly 172

concentrated bacterial suspensions at ambient conditions using a Double Ganesha 173

AIR system (SAXSLAB, Denmark). Monochromatic radiation with a wavelength of = 174

1.54 Å was generated by a rotating anode source (Cu, MicroMax 007HF, Rigaku 175

Corporation, Japan). All data sets were recorded with a position-sensitive detector 176

(PILATUS 300K, Dectris) placed at different distances from the sample to cover a 177

wide range of scattering vectors q with 0.002 Å-1 < q < 0.5 Å-1, where q is given as 178

2sin

4

qq

with representing the wavelength of the incident beam and the 179

scattering angle. 180

The 2-dimensional scattering patterns were converted into 1-dimensional intensity 181

profiles of I(q) vs. q by radial averaging and subsequently normalized to the intensity 182

of the incident beam, the sample thickness and the accumulation time. Background 183

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

9

correction was performed by subtracting the signal of a glass capillary ( = 1 mm, 184

Hilgenberg, Germany) filled with HEPES buffer. The software SasView 4.2 was used 185

for data analysis. 186

187

188

Results and Discussion 189

190

A simplified model to describe the magnetosome chain 191

As SAXS represents an indirect technique acquiring information in reciprocal space, 192

appropriate models have to be established to derive structural parameters of the 193

sample in real space. Therefore, at first a model description of the magnetosomes 194

based on both, the architecture of individual particles and their arrangement within 195

the cell had to be developed in order to allow for a detailed evaluation of the x-ray 196

scattering data. 197

One magnetosome consists of a magnetite crystal, surrounded by a proteinaceous 198

membrane (Fig. 2a), which exerts strict control over the growth of the crystals, and 199

their arrangement into well-ordered chains (4, 12). Based on previous TEM imaging 200

of magnetosome chains within M. gryphiswaldense (24, 25), magnetosome 201

biogenesis is here considered as the formation of a single linear chain composed of 202

individual core-shell particles, which exhibit sharp phase boundaries and are closely 203

packed (Fig. 2c). In the case of a strictly monodisperse system, such an initial 204

intuitive approach would result in an x-ray scattering profile (Intensity I(q) vs. q) 205

exhibiting pronounced extrema and a strong decay in scattering intensity in the 206

regime of low q values, but fails to explain the additional scattering contribution seen 207

in the high q range. 208

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

10

Therefore, an alternative model is proposed, which contains all structural parameters 209

of interest to the here presented study: the magnetosome particle radius (R), the 210

degree of polydispersity and the distance between neighbouring particles (l). For 211

simplification, no indirect Fourier transformation was included in the evaluation. 212

Based on the model presented in Fig. 2b, it can be assumed that the x-ray contrast of 213

a magnetosome is mainly determined by the electron densities of magnetite (~1500 214

electrons/nm3) and the attached proteins, for which an electron density in the order of 215

650 electrons/nm3 is estimated by grey scale analysis of TEM micrographs. 216

In this model, the magnetite crystal was regarded as a sphere (rather than a facetted 217

object) decorated densely, but randomly with smaller spheres simulating the protein-218

crowded particle coverage (Fig. 2b). Furthermore, in order to represent the biological 219

heterogeneity more accurately, a certain polydispersity with respect to the 220

magnetosome radius had to be incorporated into the theoretical description. Under 221

these restrictions, in a multi-particle model the sharp phase boundary smears out and 222

the magnetosome effectively appears as a simple homogeneous sphere with radius 223

R. It has to be considered, though, that this simplified approach does not account for 224

contributions originating from empty vesicles, invaginations and the cellular matrix, as 225

these do not feature a significant x-ray contrast at the interior. A linear chain of such 226

spheres can then be assumed to describe the geometric arrangement of 227

magnetosomes within the cell (Fig. 2c and d). However, since in a helix-type folding 228

pattern the helical pitch size, i.e. the length of one single turn, would most probably 229

exceed a dimension of 200 nm (24, 25), distinguishing between a helical and a linear 230

chain-like arrangement is beyond the resolution limit of most conventional SAXS 231

setups. 232

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

11

Summarizing the here presented model, it is suggested that a magnetosome can be 233

simplified as a homogeneous large sphere with a typical size range of 10-70 nm 234

surrounded by 3-6 nm thick proteinaceous phospholipid bilayer that is invaginated 235

from several non-specific cytoplasmic membrane locations (4, 26). Analysis of the 236

magnetosome subproteom (i.e. the complement of proteins specifically associated 237

with the magnetosome membrane) suggested an unusually crowded protein 238

composition on the membrane, and a tight packing with transmembrane domains of 239

integral proteins (27). Depending on the size and folding (number of transmembrane 240

helices and extra-magnetosomal domains) the magnetosome proteins can be 241

modelled as small objects (spheres) typically in the size range of 4-10 nm (28). For 242

simplicity, a coiled (spherical) supramolecular arrangement/folding pattern is 243

assumed. 244

In the regime of small q (corresponding to large dimensions in real space), the x-ray 245

scattering pattern is insensitive to the fine details of the protein shell. Hence, while 246

the proteins significantly increase the contrast of the shell, they are not 247

distinguishable as separate small spherical objects. Consequently, the SAXS profile 248

is dominated by the overall size of the particles, and data analysis in the lower q 249

range yields the magnetosome radius R, i.e. the outer radius determined by the 250

thickness of the magnetosome membrane (shell). However, in the regime of high q 251

(corresponding to smaller dimensions in real space), x-ray scattering detects 252

fluctuations in the electron density distribution on a smaller length scale and therefore 253

is sensitive to finer structural motives such as the proteins themselves. 254

255

256

257

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

12

Calculation of the scattering function 258

The intensity I(q) is determined by the square of the amplitude of the scattered wave, 259

which, in turn, is given by equation 1 in case of a particle with the volume Vp and an 260

excess electron density distribution Δ(r). 261

Vp

drrqr

qrrqA 2sin

4 (eq. 1). 262

For a system of monodisperse spheres with radius R, eq. 1 can be expressed as a 263

function of the particle radius according to equation 2. 264

33 cossin3

3

4

qR

qRqRqRRqA

(eq. 2). 265

In order to calculate the intensity for polydisperse systems composed of different 266

particles species with size i, the squared amplitudes qAi

2 have to be added up and 267

weighted by their respective number densitypV

N. 268

269

X-ray scattering analysis of magnetosome particles 270

The scattering intensity I(q) originating from non-interacting small and large spheres 271

is given by the corresponding amplitudes A(q) according to equation 3 (29): 272

i i ji ij

ij

jsmallismallsmallbigsmallbigqd

qdqAqA

qd

qdqAqAAqAqI

sin2

sin,,

22 (eq. 3) 273

where d denotes the distance between the centres of gravity of the large and the 274

small spheres whereas dij represents the mutual distances between small spheres. 275

The first two terms are equal to the intensities of non-interacting spheres of the 276

corresponding size (cf. eq. 2). In the case of large spheres, the first term decays 277

rapidly with I(q) ~q-4, and therefore does not contribute significantly to the scattering 278

intensity at high values of q (cf. Fig. S1), whereas the second term associated with 279

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

13

the small spheres (high amount) plays the dominant role. Due to polydispersity in 280

real-world systems, the factor

qd

qdsincancels out the third term at high q. 281

Concomitantly, the average over the fourth term vanishes as the distance dij varies 282

randomly. 283

As discussed above, SAXS represents a low-resolution method at low scattering 284

angles. As a consequence, term 2, i.e. the sum of the contribution of the small 285

spheres, can be alternatively described as an additional contribution to the electron 286

density of the entire large sphere. The protein decoration may then be regarded as 287

independent small density fluctuations and thus according to scattering theory, the 288

intensity of these fluctuations will add up to describe the scattering of a much larger 289

object with high contrast (here corresponding to the magnetosome). 290

Therefore, particularly in view of potential applications in on-line analysis of 291

magnetosome biosynthesis, the SAXS data analysis of these biogenic particles can 292

be simplified by assuming that the total intensity is modelled by a term corresponding 293

to the large spheres and an additional contribution originating from the small objects, 294

e.g. I(q)=Isphere(q) + Ismall(q) (eq. 4). Irrespective of the particle morphology, the 295

scattering intensity I(q) for a dilute solution of small objects is given by Guinier’s law 296

2

2

2

3exp q

RcMqI

g

wsmall (eq. 5), where Δ denotes the electron density 297

contrast of the small spheres, c the weight concentration, Mw the molecular weight 298

and Rg the radius of gyration. For the subsequent analysis it is important to note that 299

Ismall(q) mainly contributes to the scattering intensity for q > 0.05 Å-1 (Fig. 3) and can 300

be approximated by Guinier’s law, which allows us to perform profile fitting based on 301

a minimum of adjustable parameters. 302

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

14

With the terms for the evaluation of magnetosome and protein radius at hand, 303

anaerobically grown, magnetosome-producing cells were analysed by TEM and 304

SAXS (Fig. 3). In this experiment, the radius of the magnetosome is derived from the 305

position of the first minimum (q ≈ 0.02 Å-1) as R = 19 nm ± 17% (Gaussian size 306

distribution). The radius of gyration of the small objects (proteins) is determined as 307

1.7 nm and correlates to the radius R of a solid sphere via equation 6, 308

RRg5

3 (eq. 6), yielding R = 2.2 nm. This value is comparable to the general 309

dimensions expected for the coiled proteins based on TEM imaging (28, 30). 310

Hence, it can be concluded that within the framework of the afore-mentioned 311

assumptions, the scattering pattern is well described by the proposed model for 312

q > 0.015 Å-1. 313

To further test this concept, magnetosome-free cells of M. gryphiswaldense 314

(aerobically grown), which are characterized by the lack of magnetite crystals, but still 315

contain in some cases (empty) membrane vesicles (26) were analysed. For this 316

purpose, SAXS data obtained from magnetite-free and magnetite-producing cells 317

(Fig. 3c-e) were compared. Intriguingly, both magnetite-free systems (cf. blue and 318

green curves in Fig. 3e) do not exhibit the characteristic oscillations observed for 319

producing cells in the mid q range (q [0.08 Å-1, 0.03 Å-1]), but instead show a 320

similar linear decay at low q (q < 0.01 Å-1). In contrast, at high values of q, where 321

small nanostructural features dominate the curve, the scattering profiles of 322

magnetosome-deficient cells and magnetosome-producing bacteria (i.e. cells that 323

synthesize magnetite cores enveloped by the magnetosome membrane) are almost 324

identical. In some cases, here exemplified by the green curve, even a clear kink is 325

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

15

observed (q ≈ 0.06 Å-1), indicating the presence of a huge amount of very well-326

defined small objects such as proteins, which are highly abundant in the cells. 327

These observations strongly support the assumption that the total scattering intensity 328

of one magnetosome (form factor scattering) can be deconvoluted into partial 329

intensities assigned to large and small spherical objects, Isphere(q) and Ismall(q). Based 330

on these results one can in principle even consider the magnetosome-deficient 331

phenotype as background when investigating the scattering observed for 332

anaerobically grown M. gryphiswaldense cells which biosynthesize magnetosomes. 333

However, while there were no phenotypical differences between the two specimens 334

of magnetite-free bacteria identified by TEM (Fig. 3c,d), their corresponding SAXS 335

profiles clearly showed a deviation, particularly in the regime of high q (Fig. 3e, blue 336

and green curves), which indicates a substantial biological variability in these 337

systems, and thus complicates the identification of a suitable background curve. 338

The assembly of the magnetosomes into ordered chains is the major reason for the 339

deviation between the experimentally observed scattering intensity at low q (Fig. 3e, 340

black curve) and the here applied fit model. Due to the high degree of order (aligned 341

magnetosomes), Bragg’s law is suited to determine inter-particles separation 342

distances, such that a correlation length of approximately 47 nm can be determined 343

based on the position of the first maximum (q ≈ 0.0134 Å-1). 344

345

Determination of structural magnetosome assembly into chains 346

Chain assembly results from the coordinated interplay of collaborative magnetic 347

interactions, which occur naturally between magnetite nanocrystals, in combination 348

with active protein-mediated bio-assembly mechanisms (31, 32). Cellular structures 349

stabilize the string of magnetic dipoles, thereby avoiding a magnetic moment induced 350

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

16

collapse into rings or irregular clusters (33). In this geometric arrangement, the actin-351

like MamK filament anchors the magnetosome chain and generates an intracellular 352

compass needle to maximize the cell response to even low magnetic fields 353

experienced under environmental conditions. The pole-to-midcell treadmilling growth 354

of MamK filaments provides the key driving force for magnetosome movement (34). 355

Typically, wild type cells of M. gryphiswaldense synthesize 1-2 magnetosome chains 356

that extend across approximately two-third of the length of the cell and are located at 357

mid-cell. An average chain is composed of 20-40 particles with a mean center-to-358

center distance of about 50–60 nm (22, 24, 25). 359

According to the decoupling approximation, the total scattering intensity is 360

proportional to the product of the scattering originating from one particle (form factor) 361

and the scattering due to inter-particle interactions (structure factor). The structure 362

factor S(q) is related to the probability distribution function of inter-particle distances 363

g(r) and can be expressed according to equation 7 for isotropic systems. 364

0

2 sin141 dr

qr

qrrgrqS (eq. 7) 365

In an ideal solution featuring no position or orientation correlations, the equation 366

S(q) = 1 can be assumed, whereas ordered crystal-like particle arrangements exhibit 367

S(q) > 2.85 according to the Hansen Verlet criterion (35). The here presented 368

approach to consider the data set obtained from the magnetite-deficient bacteria as 369

background signal for the scattering of magnetosome-containing cells affords the 370

possibility to experimentally estimate the particle-particle interaction via the 371

experimental structure factor S(q), where the latter is obtained by dividing the 372

scattering intensity of the magnetite-containing sample by the intensity of the 373

respective magnetite-deficient sample. 374

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

17

Next, magnetosome biomineralization and chain formation during cell growth were 375

investigated. Therefore, samples were drawn at different production time points and 376

subsequently analysed by TEM and SAXS (Fig. 4). Using the scattering data 377

obtained at time zero as background (Fig. 4a, black line), the experimental structure 378

factors were determined as S(q) = 1 at start, S(q) ≈ 1.8 after 8 h and S(q) ≈ 3.2 after 379

23 h of biomineralization, where this corresponds to the degree of particle interaction 380

and is accompanied by the evolution from a non-ordered to a highly ordered spatial 381

correlation of the magnetite cores into a single chain. In support of the here proposed 382

theoretical model, the extracted structural parameters are in very good agreement 383

with TEM results (Fig. 4b), which clearly demonstrate the 1D crystalline order of 384

magnetosomes after 23 h. 385

Such an ordered chain-like nanoparticle array resembles a chain as described by a 386

linear pearl model (36), which is based on the form factor for NS spheres with radius 387

R linearly joined by short strings of length l (cf. Fig. 2; the thickness of the strings is 388

not taken into account). The scattering intensity I(q) for such an assembly is given by 389

equation 8. 390

2

31

2 cossin3

sin2

qR

qRqRqR

nq

nqnNNV

V

NqI

S

S

N

n

SSpearlsN

(eq. 8). 391

This consideration accounts for an increase in the scattering intensity on length 392

scales of the order of q(2R+l) ≈ 2πk, where k is an integer. 393

The scattering data in Fig. 4c were obtained from fully mature M. gryphiswaldense 394

cells after subtraction of the scattering contribution Ismall(q). For comparison, the 395

theoretical scattering curve of spheres with a radius of 19 nm (± 17%, Gaussian size 396

distribution) and the intensity profile of 5 monodisperse spheres arranged in a line 397

with a center-to-center distance of 2R+l = 50 nm are depicted. The modelled 398

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

18

interaction distance of 50 nm is in very good agreement with the oscillation pattern in 399

the scattering data arising from inter-particle interactions (first maximum observed at 400

q ≈ 0.013 Å-1), and most importantly is confirmed by TEM analysis, which reveals a 401

separation distance of 2R+l = 53 ± 6.0 nm. 402

403

404

Conclusions 405

Routinely used methods to monitor magnetosome biosynthesis such as cmag (16, 406

17), iron content measurement and EXAFS (18) as well as TEM, can often only 407

provide a semi-quantitative assessment of magnetosome structure and content within 408

the investigated cells, thus rendering them comparably unpractical for on-line process 409

control and generating a strong demand for efficient quantitative bulk methods. 410

Therefore, the present study introduces SAXS as a non-destructive (i.e. leaving 411

membranes and proteins intact) stand-alone tool for process monitoring during 412

magnetosome production. 413

In order to model the here obtained scattering profiles, magnetosomes were 414

considered as polydisperse spheres arranged into a single linear chain, which is 415

reflected in the scattering signal at low q, whereas the cell’s densely packed protein 416

background dominates the signal at high q. Based on these simplifications, a linear 417

pearl model without indirect Fourier transformation was applied to describe 418

concentrated cell suspensions grown under anaerobic (magnetosome producing) 419

and aerobic (magnetosome-free) conditions. The linear pearl model based on five 420

polydisperse spheres turned out to be best suited for describing the experimental 421

data, yielding magnetosome sizes that could be essentially confirmed by TEM 422

analysis. In conclusion, this study demonstrates that SAXS measurements performed 423

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

19

on a laboratory instrument can be used as a convenient, readily accessible, and 424

potentially even automated method to monitor bulk magnetosome biosynthesis in 425

growing cells and cultures. Most importantly, SAXS is a volume-averaging, 426

quantitative technique providing statistically accurate information on size, shape and 427

arrangement of nanostructural motives in a sample. Additionally, the lab-scale in-428

house equipment employed in this study enables at-line analysis of magnetosome 429

biomineralization and perspectively even on-line process control during fermentation. 430

Therefore, SAXS has the potential to complement and partially replace laborious and 431

time-consuming quantitative TEM analyses, thereby allowing for routine 432

magnetosome characterization. Applying the here proposed model, structural 433

parameters such as the dimensions of the inorganic crystal core and the protein-rich 434

membrane as well as the arrangement and assembly of the magnetosomes can be 435

extracted with high statistical accuracy. In view of a more general applicability to 436

different species and their magnetosome mutants grown under different conditions, 437

the model can be easily adapted and modified to account for different particle shapes 438

and arrangements within the cell e.g. by implementing form factors for non-spherical 439

scattering objects. The general contrast situation with respect to the electron density 440

heterogeneities within the sample (i.e. Fe-containing mineral vs. organics) should not 441

be affected by the specific strain of bacteria used, thus rendering the proposed 442

method a highly versatile tool for elucidating further details in magnetosome 443

formation. 444

445

446

447

448

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

20

Acknowledgments 449

We thank the Bavarian Polymer Institute (BPI) for providing access to small-angle x-450

ray scattering facilities within the Keylab Mesoscale Characterization: Scattering 451

Techniques. The authors are grateful to M. Schlotter (Dept. Microbiology, University 452

of Bayreuth) for expert technical assistance and M. Völkl for preparing some of the 453

samples as part of his Master thesis at the Dept. Microbiology and at the Chair for 454

Process Biotechnology (Prof. R. Freitag). This work benefited from the use of the 455

SasView application, originally developed under NSF award DMR-0520547. SasView 456

contains code developed with funding from the European Union's Horizon 2020 457

research and innovation programme under the SINE2020 project, grant agreement 458

No 654000. 459

460

461

Funding 462

The authors gratefully acknowledge financial support by the DFG via the 463

Collaborative Research Center SFB 840. In addition, this project has received 464

funding from the European Research Council (ERC) under the European Union’s 465

Horizon 2020 research and innovation programme (grant agreement No 692637 to 466

D.S.). 467

468

469

470

471

472

473

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

21

References 474

475

1. Lefevre C.T , Bazylinski D.A.. 2016. Ecology, Diversity, and Evolution of 476

Magnetotactic Bacteria. Microbiology and Molecular Biology Reviews 77:497-477

526. 478

479

2. Lohße A, Borg S, Raschdorf O, Kolinko I, Tompa E, Pósfai M, Faivre D, 480

Baumgartner J, Schüler D. 2014. Genetic dissection of the mamAB and mms6 481

operons reveals a gene set essential for magnetosome biogenesis in 482

Magnetospirillum gryphiswaldense. J. Bacteriol. 196:2658-2669. 483

484

3. Barber-Zucker S, Keren-Khadmy N., Zarivach R. 2016. From invagination to 485

navigation: The story of magnetosome-associated proteins in magnetotactic 486

bacteria. Protein Science. 25:338-351. 487

488

4. Uebe R, Schüler D. 2016. Magnetosome biogenesis in magnetotactic bacteria. 489

Nat. Rev. Microbiol. 14:621-637. 490

491

5. Staniland SS, Ward B, Harrison A, van der Laan G, Telling N. 2007. Rapid 492

magnetosome formation shown by real-time X-ray magnetic circular dichroism. 493

Proc. Natl. Acad. Sci. U.S.A. 104:19524-19528. 494

495

6. Staniland SS, Rawlings AE. 2016 Crystallizing the function of the 496

magnetosome membrane mineralization protein Mms6. Biochem. Soc. Trans. 497

44:883-890. 498

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

22

499

7. Hergt R, Dutz S, Röder M. 2008. Effects of size distribution on hysteresis 500

losses of magnetic nanoparticles for hyperthermia. J. Phys. Condens. Matter 501

20:385214. 502

8. Alphandéry E, Guyot F, Chebbi I. 2012. Preparation of chains of 503

magnetosomes, isolated from Magnetospirillum magneticum strain AMB-1 504

magnetotactic bacteria, yielding efficient treatment of tumors using magnetic 505

hyperthermia. Int. J. Pharm. 434:444-452. 506

507

9. Heinke D, Kraupner A, Eberbeck D, Schmidt D, Radon P, Uebe R, Schüler D, 508

Briel A. 2017. MPS and MRI efficacy of magnetosomes from wild-type and 509

mutant bacterial strains. IJMPI 3:1706004. 510

511

10. Kraupner A, Eberbeck D, Heinke D, Uebe R, Schüler D, Briel A. 2017. 512

Bacterial magnetosomes - nature's powerful contribution to MPI tracer 513

research. Nanoscale 9:5788-5793. 514

515

11. Tanaka T, Takeda H, Ueki F, Obata K, Tajima H, Takeyama H, Goda Y, 516

Fujimoto S, Matsunaga T. 2004. Rapid and sensitive detection of 17β-estradiol 517

in environmental water using automated immunoassay system with bacterial 518

magnetic particles. J. Biotechnol. 108:153-159. 519

520

12. Mickoleit F, Schüler D. 2019. Generation of nanomagnetic biocomposites by 521

genetic engineering of bacterial magnetosomes. Bioinspired Biomim. 522

Nanobiomat. 8:86-98. 523

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

23

524

13. Vargas G, Cypriano J, Correa T, Leão P, Bazylinski D, Abreu F. 2018. 525

Applications of magnetotactic bacteria, magnetosomes and magnetosome 526

crystals in biotechnology and nanotechnology: mini-review. Molecules 527

23:2438. 528

529

14. Heyen U, Schüler D. 2003. Growth and magnetosome formation by 530

microaerophilic Magnetospirillum strains in an oxygen-controlled fermentor. 531

Appl. Microbiol. Biotechnol. 61:536-544. 532

533

15. Fernández-Castané A, Li H, Thomas OR, Overton TW. 2017. Flow cytometry 534

as a rapid analytical tool to determine physiological responses to changing O2 535

and iron concentration by Magnetospirillum gryphiswaldense strain MSR-1. 536

Sci. Rep. 7:13118. 537

538

16. Schüler D, Uhl R, Bäuerlein E. 1995. A simple light scattering method to assay 539

magnetism in Magnetospirillum gryphiswaldense. FEMS Microbiol. Lett. 540

132:139-145. 541

542

17. Schüler D. 1999. Formation of magnetosomes in magnetotactic bacteria. J. 543

Mol. Microbiol. Biotechnol. 1:79-86. 544

545

18. Baumgartner J, Morin G, Menguy N, Gonzalez T P, Widdrat M, Cosmidis J, & 546

Faivre D. 2013. Magnetotactic bacteria form magnetite from a phosphate-rich 547

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

24

ferric hydroxide via nanometric ferric (oxyhydr) oxide intermediates. PNAS, 548

110:14883-14888. 549

550

19. Fdez-Gubieda ML, Muela A, Alonso J, García-Prieto A, Olivi L, Fernández-551

Pacheco R, Barandiarán JM. 2013. Magnetite biomineralization in 552

Magnetospirillum gryphiswaldense: time-resolved magnetic and structural 553

studies. ACS nano 7:3297-3305. 554

555

20. Krueger S, Olson GJ, Rhyne JJ, Blakemore RP, Gorby YA, Blakemore N. 556

1989. Small angle neutron and X-ray scattering from magnetite crystals in 557

magnetotactic bacteria. J. Magn. Magn. Mater. 82:17-28. 558

559

21. Hoell A, Wiedenmann A, Heyen U, Schüler D. 2004. Nanostructure and field-560

induced arrangement of magnetosomes studied by SANSPOL. PHYSICA B 561

350:E309-E313. 562

563

22. Orue I, Marcano L, Bender P, García-Prieto A, Valencia S, Mawass MA, Gil-564

Cartón D, Alba Venero D, Honecker D, García-Arribas A, Fernández Barquín 565

L, Muela A, Fdez-Gubieda ML. 2018. Configuration of the magnetosome 566

chain: a natural magnetic nanoarchitecture. Nanoscale 10:7407-7419. 567

568

23. Molcan M, Petrenko VI, Avdeev MV, Ivankov OI, Garamus VM, Skumiel A, 569

Jozefczak A, Kubovcikova M, Kopcansky P, Timko M. 2017. Structure 570

characterization of the magnetosome solutions for hyperthermia study. J. Mol. 571

Liq. 235:11-16. 572

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

25

573

24. Katzmann E, Müller FD, Lang C, Messerer M, Winklhofer M, Plitzko JM, 574

Schüler D. 2011. Magnetosome chains are recruited to cellular division sites 575

and split by asymmetric septation. Mol. Microbiol. 82:1316-1329. 576

577

25. Katzmann E, Eibauer M, Lin W, Pan Y, Plitzko JM, Schüler D. 2013. Analysis 578

of magnetosome chains in magnetotactic bacteria by magnetic measurements 579

and automated image analysis of electron micrographs. Appl. Environ. 580

Microbiol. 79:7755-7762. 581

582

26. Raschdorf O, Forstner Y, Kolinko I, Uebe R, Plitzko JM, Schüler D. 2016. 583

Genetic and ultrastructural analysis reveals the key players and initial steps of 584

bacterial magnetosome membrane biogenesis. PLoS Genet. 12:e1006101. 585

586

27. Raschdorf O, Bonn F, Zeytuni N, Zarivach R, Becher D, Schüler D. 2018. A 587

quantitative assessment of the membrane-integral sub-proteome of a bacterial 588

magnetic organelle. J. Proteomics 172:89-99. 589

590

28. Nudelman H, Zarivach R. 2014. Structure prediction of magnetosome-591

associated proteins. Front. Microbiol. 5:9. 592

593

29. Glatter O, Kratky O. 1982. Small angle X-ray scattering. Academic press, 594

London 595

596

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

26

30. Erickson HP. 2009. Size and shape of protein molecules at the nanometer 597

level determined by sedimentation, gel filtration, and electron microscopy. Biol. 598

Proced. Online, 11:32. 599

600

31. Klumpp S, Faivre D. 2012. Interplay of magnetic interactions and active 601

movements in the formation of magnetosome chains. PLoS ONE 7:e33562. 602

603

32. Ghaisari S, Winklhofer M, Strauch P, Klumpp S, Faivre D. 2017. 604

Magnetosome organization in magnetotactic bacteria unraveled by 605

ferromagnetic resonance spectroscopy. Biophys J. 113:637-644. 606

607

33. Bennet M, Bertinetti L, Neely RK, Schertel A, Körnig A, Flors C, Müller FD, 608

Schüler D, Klumpp S, Faivre D. 2015. Biologically controlled synthesis and 609

assembly of magnetite nanoparticles. Faraday discuss. 181:71-83. 610

611

34. Toro-Nahuelpan M, Müller FD, Klumpp S, Plitzko JM, Bramkamp M, Schüler 612

D. 2016. Segregation of prokaryotic magnetosomes organelles is driven by 613

treadmilling of a dynamic actin-like MamK filament. BMC Biol. 14:88. 614

615

35. Hansen JP, Verlet L. 1969. Phase transitions of the Lennard-Jones system. 616

Phys. Rev. 184:151-161. 617

618

36. Dobrynin AV, Rubinstein M, Obukhov SP. 1996. Cascade of transitions of 619

polyelectrolytes in poor solvents. Macromol. 29:2974-2979. 620

621

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

27

Figure Captions 622

623

Fig. 1 Magnetosome arrangement in M. gryphiswaldense. TEM micrograph of wild type cells 624

illustrating the chain-like particle arrangement (a), as well as a sketch of a single cell with a schematic 625

of a magnetosome consisting of one magnetite core and its surrounding magnetosome membrane (b). 626

The latter contains phospholipids and about 30 different Mam proteins with various functions in control 627

of magnetosome biomineralisation. (4) 628

629

Fig. 2 Form factor model of magnetosomes. (a) TEM micrograph of isolated magnetosomes from 630

M. gryphiswaldense (Mickoleit, unpublished data), exhibiting the characteristic core-shell structure. (b) 631

Sketch of a single magnetosome (as shown in Fig. 1) and the corresponding x-ray contrast model. 632

SAXS is sensitive to the different electron density distributions of the core and the surrounding shell, 633

which mainly consist of magnetite and proteins, respectively. Both contributions are modelled as 634

spheres of different sizes. The resulting contrasts contribute in a characteristic way to specific regions 635

in the scattering data, which consequently contain information on both the overall particle radius R as 636

well as the radius of gyration Rg of the individual proteins. (c) Chain-like arrangement of 637

magnetosomes within one cell as seen by TEM. Separation distances between single particles are 638

indicated by arrows, while the membranes surrounding the magnetite cores are highlighted by dashed 639

lines. (d) Model of the particle chain shown in (c). SAXS measurements detect the center-to-center 640

distances of two magnetosomes given by 2R + l. 641

642

Fig. 3 Comparison of magnetosome-free and producing cells of M. gryphiswaldense. (a) TEM 643

micrograph of representative magnetosome-producing cells grown under anaerobic conditions. (b) 644

Scattering patterns obtained from cell suspensions (blue circles) shown in (a). Additionally, the 645

calculated scattering patterns of small objects according to eq. 5 (solid black line) and polydisperse 646

spheres according to eq. 2 are shown (red line). The spheres exhibit a radius of 19 nm with a 647

Gaussian distribution with standard deviation of 17%. (c/d) TEM micrographs of representative, 648

magnetite-free cells of M. gryphiswaldense. (e) Scattering patterns of the corresponding 649

M. gryphiswaldense cell suspensions. The curves are presented with a vertical shift for better visibility: 650

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

28

Open symbols, magnetite-producing M. gryphiswaldense cells as shown in (a); filled symbols, 651

magnetite-free cells (green or blue symbols correspond to cells shown in (c) or (d), respectively). 652

Additionally, the calculated scattering profile of the small particles is given for comparison (grey line). 653

654

655

Fig. 4 Scattering patterns and TEMs of M. gryphiswaldense at different sampling times during 656

cultivation. (a) Patterns at production start (t = 0 h, black line, no magnetite due to growth conditions), 657

after 8 h (red squares) and 23 h (blue circles), and (b) corresponding TEM micrographs. (c) The 658

scattering of fully matured cells (symbols, after subtraction of the scattering contribution Ismall(q)) are 659

compared to the form factor scattering of polydisperse, non-interacting spheres (thin blue line, R = 19 660

± 3 nm, Gaussian distribution) and the scattering of 5 monodisperse spheres arranged in a linear 661

chain (thin red line, R = 19 nm, l = 12 nm). A model allowing for polydispersity in the system is 662

represented by the green line. 663

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from

on October 31, 2020 by guest

http://aem.asm

.org/D

ownloaded from