OUP user menu

Interactive optical trapping shows that confinement is a determinant of growth in a mixed yeast culture

Nils Arneborg , Henrik Siegumfeldt , Grith H. Andersen , Peter Nissen , Vincent R. Daria , Peter John Rodrigo , Jesper Glückstad
DOI: http://dx.doi.org/10.1016/j.femsle.2005.03.008 155-159 First published online: 1 April 2005


Applying a newly developed user-interactive optical trapping system, we controllably surrounded individual cells of one yeast species, Hanseniaspora uvarum, with viable cells of another yeast species, Saccharomyces cerevisiae, thus creating a confinement of the former. Growth of surrounded and non-surrounded H. uvarum cells was followed under a microscope by determining their generation time. The average generation time of surrounded H. uvarum cells was 15% higher than that of non-surrounded cells, thereby showing that the confinement imposed by viable S. cerevisiae cells on H. uvarum inhibits growth of the latter. This study is the first to demonstrate that confinement is a determinant of growth in a microbial ecosystem.

  • Interactive optical trapping
  • Confinement
  • Growth inhibition
  • Yeast

1 Introduction

We have previously demonstrated that viable cells of the yeast Saccharomyces cerevisiae at a high density cause growth arrest of two non-Saccharomyces yeast species in mixed cultures [1,2]. This finding made us speculate whether viable S. cerevisiae cells are able to affect growth of non-Saccharomyces yeast cells in a mixed culture by simply surrounding them; i.e., whether the former can impose a confinement stress on the latter. In mixed cultures with S. cerevisiae, however, non-Saccharomyces yeasts may be exposed to several growth-inhibiting factors, including nutrient limitation (e.g., carbon and nitrogen sources and oxygen) and growth-inhibitory compounds (e.g., ethanol and killer toxins [3]), as a result of the growth and metabolism of S. cerevisiae.

Accordingly, verification of our confinement-stress hypothesis requires an experimental approach that should clearly separate the effect of confinement from the effects of the other factors mentioned above on growth of non-Saccharomyces yeast. That is, first, it should ensure non-Saccharomyces yeast cells to be controllably surrounded by viable S. cerevisiae cells. Second, it should uncouple the effect of high-cell-density growth of S. cerevisiae from the effect of its metabolism on growth of a non-Saccharomyces yeast while still exposing all cells to identical nutrient and metabolite concentrations. However, fulfilment of these two requirements in combination has not been possible with existing experimental set-ups.

Conventional optical traps, commonly known as optical tweezers, consist of a single, highly focused laser beam that requires immersion objectives with high numerical aperture (NA) to obtain only one trap [4]. In contrast, a recently proposed user-interactive optical trapping system [5] can generate a specified number of traps where size, shape and position of the individual traps can be adjusted by the user during the experiment via a graphical user interface. Furthermore, high-NA immersion objectives are not a prerequisite, which allows for longer working distance objectives that facilitates imaging in a multitude of standard lab-ware. This novel system enables real-time interactive optical manipulation of multiple, growing cells without affecting their generation times [5]. Here, we apply this system on a mixed culture of the two facultatively fermentative yeasts, S. cerevisiae and Hanseniaspora uvarum, and we provide evidence that a confinement stress imposed by viable S. cerevisiae cells on H. uvarum inhibits growth of the latter.

2 Materials and methods

2.1 Micro-scale, mixed-culture growth experiments

2.1.1 Yeast strains and media

H. uvarum CBS 314 and S. cerevisiae (Saint Georges S101, Bio Springer, France) were used. Small aliquots of culture broth from 22-h-old single cultures of H. uvarum and S. cerevisiae, grown with agitation (140 rpm) in rich YPD medium (9 g l−1 glucose, pH 5.6) at 25 °C, were transferred to a perfusion chamber (CoverWell, Sigma-Aldrich) with fresh YPD medium, and the cells were allowed to settle on the glass cover-slip.

2.1.2 Microscopy and interactive optical trapping

The chamber was sealed with grease and subsequently mounted on an inverted microscope (DM-IRB, Leica) with a non-immersion objective lens (63×, NA = 0.75; C PLAN, Leica). Interactive optical trapping system based on the generalized phase contrast method [6,7] allows synthesis of real-time reconfigurable light intensity patterns that are brought to the sample chamber through the microscope fluorescence port [8,9]. Fig. 1 shows the diagram of the system for generating arbitrary arrays of optical traps for multiple-cell manipulation. An incident light from an expanded and collimated Titanium:Sapphire continuous-wave laser source (3900S, Spectra Physics), operating at a wavelength of 830 nm, is dynamically encoded with spatial phase patterns using a computer-programmable phase-only spatial light modulator (PPM, Hamamtsu). The succeeding components in the optical train facilitate efficient transformation of the phase pattern into a high-contrast intensity distribution that can be used as optical traps for manipulating cells. Using a graphical computer interface, the user can interactively encode the phase patterns that relates to the desired number, size and shape of the optical traps. Cells, being dielectrics of higher refractive index relative to the YPD medium, get attracted to the array of optical traps. User-defined optical traps were used to arrange and maintain the S. cerevisiae cells in a ring configuration centred about a H. uvarum cell and also to locally isolate such region of interest from other potentially interfering cells. It should be stressed that the laser wavelength of 830 nm has very low absorption of water and biological molecules and therefore causes minimal photo damage of the cells [10].


The system for generating arbitrary arrays of optical traps for simultaneous manipulation of multiple cells. A phase-only spatial light modulator is used to encode the phase of an incident laser beam to generate the desired optical trapping configuration. The user can interactively manipulate cells while directly observing the cells under the microscope.

2.1.3 Calculation of generation time

Image frames were extracted from real-time recorded videos of the growth experiments with frame-to-frame time intervals of 1.7 min. Division of two H. uvarum cells was detected by the two cells exhibiting a distinct break during each cytokinesis (Fig. 3 (c)–(h)). Generation time of a given cell was calculated as the time difference between two successive cytokinetic events; e.g., time (Fig. 3 (f)) − time (Fig. 3 (d)) = 186.0 min − 116.2 min = 69.8 min. It should be noted that the distinct cell break at cytokinesis could always be observed in two successive image frames (see, e.g., Fig. 3 (c) and (d)), thus rendering the maximum inaccuracy of each generation time measurement to be 2 × 1.7 min = 3.4 min. Furthermore, due to the number of cells within a given location being too large for proper analysis after three generations (data not shown), we could only track growth of the cells for this number of generations.


Micro-scale, mixed-culture fermentation of Hanseniaspora uvarum and Saccharomyces cerevisiae. Images from one representative experiment were recorded at (a) 5.0 min; (b) 51.5 min; (c) 114.5 min; (d) 116.2 min; (e) 184.3 min; (f) 186.0 min; (g) 217.5 min; (h) 219.2 min. S. cerevisiae cells (white arrow) were trapped in a ring-shaped optical beam pattern (full line circle), surrounding an individual H. uvarum cell (dashed black arrow). This region of interest was kept free of any interference with other cells by a larger ring-shaped optical beam pattern (dashed line circle). Black arrows show cytokinesis of individual, surrounded H. uvarum cells at 114.5–116.2 min ((c) and (d)), 184.3–186.0 min ((e) and (f)), and 217.5–219.2 min ((g) and (h)) of fermentation. Non-surrounded H. uvarum cells are located at the top centre of the images. Scale bar represents 10 μm.

2.2 Statistical analysis of data

Twenty growth experiments were performed, and all cells from these experiments were analyzed (?164 cells, Table 1). Data were normally distributed (data not shown), and statistical analysis of data was performed using the two-sample t-test, assuming unequal variances and considering both sides of the distribution (two-tailed distribution).

View this table:

Growth characteristics of Hanseniaspora uvarum cells from micro-scale, mixed-culture fermentations

Cell typenAveragegt (min)SEMgt (min)Mingt (min)Maxgt (min)
  • gt, generation time.

3 Results and discussion

Using our experimental set-up, multiple S. cerevisiae and H. uvarum cells in a microscopy sample chamber can be non-invasively manipulated and brought to a region of interest (Fig. 2). This allows us to conduct controlled and repeatable micro-scale, mixed-culture growth experiments. Noticeably, neither mechanical micro-manipulation systems nor conventional optical traps can be used for this kind of experiments, as the positions of cells, and thereby interactions between several particular cells, cannot be controlled.


Preparing the Hanseniaspora uvarum and Saccharomyces cerevisiae cells for micro-scale, mixed-culture fermentation applying the user-interactive optical trapping system. Dynamically reconfigurable optical traps simultaneously bring S. cerevisiae cells to surround an individual H. uvarum cell. Arrows show the direction of movement. Images from one representative experiment were recorded at (a) 0 s; (b) 30 s; (c) 60 s; and (d) 90 s. Scale bar represents 10 μm.

Moreover, all cells within a microscopic image frame will be exposed to the same solute concentrations throughout a growth experiment, since distances between cells are mainly less than 100 μm. According to Fick's second law of diffusion, a solute with a typical diffusion coefficient of 1 × 10−6 cm2 s−1 will travel a root mean square distance of this order (?100 μm) within 1 min, which is considerably less than the duration of a typical growth experiment.

Thus, to determine whether S. cerevisiae is able to impose a growth-inhibiting, confinement stress on a non-Saccharomyces yeast, we performed micro-scale, mixed-culture growth experiments with S. cerevisiae and the non-Saccharomyces yeast, H. uvarum, by combining the use of video microscopy with the interactive optical trapping system. H. uvarum was selected due to its special budding feature (bipolar budding), making it easily distinguishable from S. cerevisiae (multilateral budding) in the microscope. Images recorded from a typical experiment are shown in Fig. 3. Throughout the growth experiments we were able, within a user-defined region of interest, to controllably surround individual H. uvarum cells with S. cerevisiae cells, thus creating a two-dimensional confinement of the former, by trapping several (?6–7) S. cerevisiae cells in a real-time adjustable optical beam pattern (Fig. 3). We were able to keep this region free of any interference from other cells by making a boundary in form of a larger ring-shaped optical beam pattern, trapping any potentially interfering cells (Fig. 3 (e)–(h)). Furthermore, we were able to follow concurrently the growth of non-surrounded H. uvarum cells within the same image frame (Fig. 3). As mentioned above, this is an advantage of our experimental approach, since both surrounded and non-surrounded H. uvarum cells within an image frame are exposed to the same concentration of nutrients and metabolites. Moreover, any effect of the optical beams (with an order of 10 mW total power in the inner beam pattern) on the growth of H. uvarum can be excluded, since the surrounded H. uvarum cells are not directly illuminated by the trapping beams, and since the average generation times of those non-surrounded H. uvarum cells trapped in the beams and those untrapped are statistically indistinguishable (P= 0.16, data not shown). Finally, any effect of local nutrient competition on the growth of H. uvarum can be neglected, since the average generation times of both surrounded and non-surrounded cells do not increase from first to third generation (data not shown). This indicates that excess nutrients are available for all H. uvarum cells throughout the growth experiments.

Our study for the first time provides evidence that a confinement stress imposed by viable cells of one yeast species, S. cerevisiae, on another yeast species, H. uvarum, inhibits growth of the latter. That is, the surrounded H. uvarum cells divide significantly slower (?15%, P= 1.6× 10−5) than the non-surrounded cells (Table 1). The molecular mechanisms explaining this phenomenon remain to be found. We have, however, previously suggested that growth inhibition of other non-Saccharomyces yeasts by high densities of viable S. cerevisiae cells is mediated by cell–cell contact [1]. This may also be the case in the current study.

Space is considered a typical object of competition among species in a variety of biological ecosystems, e.g., forests, grasslands, coral reefs, and the marine benthic zone [11]. In microbial ecosystems, however, no reports exist on spatial constraints being a determinant of growth. It may be hypothesized that the physical available space of the surrounded H. uvarum cells in our study would be restricted, and that physical space limitation therefore would play a role in the growth inhibition of these cells. This issue, however, needs further investigation.

The average generation time of surrounded (91.7 min) and non-surrounded (77.9 min) cells (Table 1), corresponding to specific growth rates of 0.45 and 0.53 h−1, respectively, are comparable to population-averaged data reported in literature on aerobically grown H. uvarum[12]. Together with the fact that H. uvarum grows much slower under strict-anaerobic conditions [13], these results suggest that oxygen is available to the H. uvarum cells in the sealed sample chamber.

Our results reveal large variations in the generation times of individual cells within the population of H. uvarum (Table 1); the coefficients of variation (CV), defined as the standard deviation divided by the mean value, being 24% and 18% for the surrounded and non-surrounded H. uvarum cells, respectively. Similar results were recently reported for prokaryotic Escherichia coli cells with a CV of 33%[14] and may be explained by the fact that regulation of the prokaryotic cell cycle is not perfect [15]. In [14], it was further concluded that eukaryotic cells are able to control their cell cycle more tightly than prokaryotic cells, since the generation times of individual eukaryotic Chlamydomonas cells had a CV of less than 5%[14]. However, the average generation time of Chlamydomonas (?600 min) [14] is an order of magnitude higher than the one of our eukaryotic cultures. Large cell-to-cell variation in generation times may therefore simply be a consequence of cells having low generation times, regardless of their phylogenetic domain.

4 Conclusion

Our study is the first to show that confinement is a determinant of growth in a microbial ecosystem. Given that crowding of yeast cells at a surface is a special feature of this work, our finding may contribute to a better understanding of cell growth regulation in, e.g., biofilms in medical, environmental and industrial settings, as well as probiotic and pathogenic communities in the gastrointestinal tract.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
View Abstract