From micro to nano: recent advances in high-resolution microscopy Yuval Garini, Bart J Vermolen and Ian T Young Improving the spatial resolution of optical microscopes is important for a vast number of applications in the life sciences. Optical microscopy allows intact samples and living cells to be studied in their natural environment, tasks that are not possible with other microscopy methods (e.g. electron microscopy). Major advances in the past two decades have significantly improved microscope resolution. By using interference and structured light methods microscope resolution has been improved to
100 nm, and with non-linear methods a ten times improvement has been demonstrated to a current resolution limit of
30 nm. These methods bring together old theoretical concepts such as interference with novel non-linear methods that improve spatial resolution beyond the limits that were previously assumed to be unreachable. Addresses Delft University of Technology, Lorentzweg 1, 2628 CJ, Delft, The Netherlands Corresponding author: Garini, Yuval ([email protected])

Current Opinion in Biotechnology 2005, 16:3–12 This review comes from a themed issue on Analytical biotechnology Edited by Keith Wood and Dieter Klaubert Available online 21st January 2005 0958-1669/$ – see front matter # 2005 Elsevier Ltd. All rights reserved. DOI 10.1016/j.copbio.2005.01.003

Abbreviations CCD charge-coupled device HELM harmonic excitation light microscopy MPM multiphoton microscopy NSOM near-field scanning optical microscopy PSF point spread function SHG second harmonic generation SPIM selective plane illumination microscopy SPR surface plasmon resonance SSIM saturated structured illumination microscopy STED stimulated emission depletion TIRF total internal reflection fluorescence

Introduction Optical microscopy is important for an immense number of applications in the life sciences. It allows one to work with intact samples including living cells and to see samples with the naked eye, advantages that are not found in other methods such as electron microscopy. Among the major developments in optical microscopy in the past century, fluorescence microscopy is probably www.sciencedirect.com

the most enabling and has become the method of choice in the majority of life-science applications. With these developments and with the natural evolution of biological studies from whole species to the molecular level, spatial resolution and the ability to distinguish fine detail has become a critical issue. The wave-like nature of light imposes a seemingly fundamental limit on the resolving power of a microscope. Even without a rigorous description, it is clear that spatial resolution is limited to approximately half the wavelength of light, approximately 200 nm for visible light (400– 750 nm). Improving this limit is a source of continuing research, with major successes in the past two decades. We have recently seen exciting developments that achieve a 10 improvement in optical resolution to
30 nm. Nevertheless, it may take time (and money) before these methods migrate from the physics laboratory to the biology laboratory. In this review we will first define resolution and then describe various techniques for modern high-resolution microscopy. Our emphasis will be on recent developments that have greatly exceeded the diffraction limit of light optics.

Resolution What is resolution anyway? Even when a very small object (say one nanometer in diameter) is observed with a microscope, its image is significantly broadened compared with the original object. First described by Abbe [1], this phenomena is a result of the diffraction of light and depends on the wavelength and the finite size of the objective lenses of the microscope. The intensity distribution of the image of a very small object is called the point spread function (PSF) (Figure 1). The PSF usually has a lateral (in-plane) radial symmetry (Figure 1a) and a larger axial broadening (along the optical axis). Figure 2 shows the intensity distribution as observed from a ‘sideview’ relative to the plane of the object. Based on the commonly used Rayleigh criterion, the minimal distance of two points that can still be resolved is approximately equal to the width of the PSF (Figure 1). According to the Rayleigh criteria, the resolution limit of a conventional microscope is given by Lateral ðin-planeÞ: dx;y ¼

0:61l NA

2l NA2 where l is the wavelength of light and NA is the numerical aperture of the lens, which is the sine of half of the Axial ðalong the optical axisÞ: dz ¼

Current Opinion in Biotechnology 2005, 16:3–12

4 Analytical biotechnology

Figure 1

(a) 1.2 20–30% of peak intensity

1.0

Intensity

0.8 0.6 0.4 0.2 0.0 –1000 –800 –600 –400 –200 0 200 400 600 x (nm) (NA = 1, λ = 500 nm) –12 –10 –8

–6

–4

–2

0 2 4 v = x*2πNA/λ

6

8

800

1000 1200

10

12

14

16

(b) 1.2

1.0

Intensity

0.8

0.6

0.4

0.2

0.0 –2000

–30

–1000

–20

–10

0 1000 z (nm) (NA = 1, λ = 500 nm) 0 10 u = z*2πNA2/λ

2000

20

3000

30

40

Current Opinion in Biotechnology

The intensity of a small object (diameter much smaller than l) as viewed through an optical microscope, also called the point spread function (PSF). (a) The intensity of a small object as observed in the plane (solid line). To distinguish two small objects from one another, the distance between them should be approximately as plotted (see the distance between the solid and dashed blue lines). The total intensity is the sum (red line) that has been artificially raised a little. It is common to use the Rayleigh criteria for the resolution. With this criterion, the sum intensity of two close objects (red line) should have an intensity minimum that is 20–27% lower than the peak intensity. (b) The same graphs shown along the optical axis z. Note that the resolution along the optical (z) axis is worse than the resolution in the lateral (x,y) plane. Two axes are shown for each graph. The upper one shows the distance for a numerical aperture (NA) of 1.0 and a wavelength of 500 nm. The lower axis is drawn for a general case and the resolution can be calculated with it for any given value of NA and l.

Current Opinion in Biotechnology 2005, 16:3–12

www.sciencedirect.com

From micro to nano: recent advances in high-resolution microscopy Garini, Vermolen and Young 5

Figure 2

(a)

Conventional

(b)

Confocal

1000

z(nm)

500

0

–500

–1000 –500

–250

0

250

500 –500 x or y (nm)

–250

0

250

500

Current Opinion in Biotechnology

The PSF as observed along the y axis for (a) a conventional microscope and (b) a confocal microscope. The intensity is shown along the z axis and along any axis in the xy plane (it has a circular symmetry in the plane). This function is calculated for NA = 1.0 and l = 500 nm. Note that in both methods the main spot is much larger along the z axis (optical axis) than in the lateral plane.

observation angle of the lens multiplied by the index of refraction, n, of the material that is in between the lens and the sample. These limits, known as the diffraction limit, form the starting point for the struggle to improve the resolution; the narrower the PSF of a system, the better the resolution that can be achieved.

High-resolution three-dimensional methods It is fascinating to see how spatial resolution has gone through evolution and revolution. We divide the existing high-resolution microscopy methods into four categories: conventional and confocal microscopy; interference methods; non-linear methods; and surface methods. The classification of a method into a category is not always well defined, but we found it easier to conceptually understand the similarities and differences between the methods using this approach. The ‘methods map’ (Figure 3) provides a general overview and the categories and methods are described below.

Conventional and confocal microscopy Conventional microscopy refers to the compound microscope, principally composed of an objective lens, an eyepiece lens, and (for infinity-corrected microscopes) www.sciencedirect.com

a tube lens. This setup has been used for 350 years since the development of the compound microscope by Hook and his outstanding 1664 publication of Micrographia [2]. Main improvements in the optical resolution over the course of the past centuries have resulted from the improved quality of the elements, most importantly the correction of aberrations in the objective lens. There are also works that have demonstrated the correction of aberrations by using adaptive optics [3,4]. These developments are important in light of the fact that almost all high-resolution methods are based on a conventional microscope. Conventional microscopy and deconvolution

With the development of digital imaging and image processing, including deconvolution methods [5], it has become possible to process two- and three-dimensional images that have been acquired with a conventional microscope and to improve the spatial resolution [6]. Most of the digital images that are measured through a microscope use a charged-coupled device (CCD). These detectors have high efficiency, collect the light from the whole image simultaneously, have a large dynamic range and long possible exposure times. This allows for the measurement Current Opinion in Biotechnology 2005, 16:3–12

6 Analytical biotechnology

Figure 3

Optical microscopy Surface methods

Three-dimensional methods Conventional

NSOM

TIRF

SPR Confocal

Non-linear

Deconvolution Interference

Multiphoton

Reversible saturation

SPIM

ApoTome SHG

CARS

I5M

HELM

2P,3P

SSIM

STED

4Pi

MMM – 4Pi Current Opinion in Biotechnology

Methods map of the different approaches for high-resolution microscopy (see text for details). Note that some of the novel methods combine more than one technique.

of high signal-to-noise ratio images, which are essential for the acquisition of high-resolution data. In some cases (e.g. thin samples with a small number of fluorescing objects), this method provides results that are as good as those obtained with a confocal microscope (Figure 4).

is the most well-established and wide-spread method in use at present. The PSF for a confocal microscope is approximately the square of the PSF of a conventional microscope and the pffiffiffi resolution improves by
2. Even more importantly, it provides good isolation of out-of-focus fluorescence. This aspect is not obvious from observing the PSF but is described, as an example, by Wilson [8].

Confocal microscopy

Confocal fluorescence microscopy [7,8] was one of the earliest methods developed for improved resolution and Figure 4

Conventional (wide field)

3000 2000

Conventional and deconvolution

3000 2000 1000

1000

) nm

)

4000 3000 2000 1000 3000 4000 2000 1000 x (nm)

y(

nm

y(

4000 3000 2000 1000

4000 z (nm)

z (nm)

4000

3000 1000 2000 x (nm)

4000

Current Opinion in Biotechnology

The deconvolution effect. (a) Raw data and (b) deconvolved data of a small part of the nucleus where telomeres have been labeled with Cy3. Fluorescence images were measured with Axioplan 2 (Zeiss) with a Planapo 63/1.4 oil immersion lens. In the raw image, some of the telomeres cannot be resolved (touching), but in the deconvolved image the telomeres are completely resolved. Deconvolution method is based on [5]. Current Opinion in Biotechnology 2005, 16:3–12

www.sciencedirect.com

From micro to nano: recent advances in high-resolution microscopy Garini, Vermolen and Young 7

Figure 5

Agarose

Tube lens

at the front and rear of the sample. This also increases the numerical aperture of the setup and results in an improved resolution. A similar interference effect can be achieved when the emitted fluorescence is collected from both objectives and combined to interfere on the CCD plane. Interference methods are usually limited to thin samples, because the quality of the interference pattern is reduced when the light passes through the sample.

Filter Sample Objective Laser Cylindrical lens Schematic diagram of a selective plane illumination microscope (SPIM). The sample is embedded in a cylinder of agarose that is held in a mechanical translation and rotation stage. The sample is excited through a thin glass window, while the objective lens observes the sample in a perpendicular axis relative to the excitation [10]. (Figure reproduced with kind permission of K Greger and EHK Stelzer.)

Confocal imaging is achieved by a single point illumination with a laser and detection through a pinhole, followed by a raster scan of the entire image. This indicates the weakness of confocal microscopy, namely the long time it takes to measure an image with a good signal-to-noise ratio. The scanning time can be reduced, for example, by using a spinning-disk that has many holes [9], but the resolution is somewhat reduced. Selective plane illumination microscopy

Selective plane illumination microscopy (SPIM) is a newly developed method that allows one to measure large-size specimens (up to few millimeters) with an improved semiconfocal illumination [10]. The highefficiency of this method, relative to confocal microscopy, results from the illumination of a whole image plane instead of a single point. SPIM permits the rapid capture of three-dimensional images so that transient biological phenomena can be detected. Its ability to detect dynamic processes in a large specimen with a resolution of
1 mm is of major importance (Figure 5).

Interference and structured illumination methods Light waves can interfere, as was shown for the first time in 1801 in Young’s wave experiments. Interference of two or more light sources can result in a periodic pattern of light on the sample (object) plane. When these patterns are used to excite fluorescence, they interact with the sample structure and the recorded emission carries higher resolution information than can be achieved by conventional microscopy. Usually, two objective lenses are used www.sciencedirect.com

Interference methods were initiated a decade ago by Lanni and colleagues [11,12] and development has continued with the introduction of the I5M, HELM and 4Pi microscopy methods [13]. A method based upon the work of Wilson and colleagues [14] that uses structured light illumination has also recently been introduced to the market (Apotome from Carl Zeiss, Gottingen, Germany). It provides the same axial resolution as a confocal microscope, while providing a better signal-to-noise ratio using ‘only’ a conventional microscope. 4Pi microscopy

4Pi microscopy [15] is based on the interference principle. It uses laser light for illumination in a confocal mode. When image restoration is added to the threedimensional scanned data [16], the resolution can be improved to the 100 nm range. Through further improvements, the acquisition time has been shortened by using many beams in parallel and a two-photon configuration [17]. This method — termed by Hell and colleagues multifocal, multiphoton 4Pi confocal microscopy (MMM-4Pi) — overcomes the long acquisition time of a single point system and allows one to measure very fine details of organelles such as mitochondria [17] and the Golgi apparatus [18] in living cells with a resolution of
100 nm (Figure 6). The 4Pi microscopy method has recently been commercialized as an extension to a confocal/two-photon microscope (Leica, Mannheim, Germany). Image interference microscopy

Image interference microscopy (I2M) uses two, high numerical aperture objective lenses and beam splitters to collect fluorescence images from the same focal plane and to let them interfere on a CCD plane [19,20]. These objective lenses can also be used to illuminate the sample from both sides with an incoherent light source (such as a mercury lamp). When used in this way, the method is called incoherent, interference, illumination microscopy (I3M) and results in an interference pattern on the sample. Gustaffson and colleagues, who first developed I3M, also combined this approach with I2M to form I5M [21]. During the measurement, a set of images is collected by scanning the sample through the system focal plane and the data are appropriately deconvolved to provide Current Opinion in Biotechnology 2005, 16:3–12

8 Analytical biotechnology

Figure 6

Mitochondrial matrix of live Saccharomyces cerevisiae labeled with green fluorescent protein and recorded with MMM-4Pi. The organelle displays strong tubular ramification of a single large body that is exclusively located beneath the plasma membrane, which is counter-stained in blue. The length of each arrow corresponds to 1 mm [17]. (Figure reproduced from [17] with kind permission of SW Hell).

high-resolution three-dimensional information. I5M provides a resolution in the range of 100 nm. Harmonic excitation light microscopy

In harmonic excitation light microscopy (HELM), four beams are combined to interfere in the sample creating high-frequency interference lines along the two main axes of the sample [22]. The system uses a laser beam coupled to the sample through a set of beam splitters and a glass block. The fluorescence is measured with a high numerical aperture objective lens on the other side of the sample. Five images are recorded with a CCD camera for different positions of the pattern on the sample, and processed to provide high-resolution images. HELM provides lateral resolution of
100 nm; it has also been shown that both axial and lateral resolution of
100 nm can be achieved [23].

Non-linear phenomena can reach very small fluorescing volumes, which results in improved spatial resolution. We believe that these methods carry the highest potential for the future, especially when combined with some of the other methods mentioned above. Multiphoton microscopy

A major improvement for studying both living cells and thick samples was achieved with the development of multiphoton microscopy (MPM) [24]. In this non-linear method, each fluorochrome is usually excited by two photons (sometimes three) such that the total energy is equal to the excited fluorochrome energy. This process is only effective at the center of the focused beam in the sample where the photon density is high, but the resolution is subsequently reduced because photons with a wavelength of 2l (or 3l) are required. This can be improved to a certain extent by using a confocal pinhole.

Non-linear methods The methods described above try to reach high resolution by improving the PSF either directly (e.g. by increasing the numerical aperture) or indirectly by using interference phenomena followed by image processing. Non-linear methods are based on a different approach. In one non-linear method (multiphoton microscopy), the fluorochromes are excited only when absorbing more than one photon. In another method (reversible saturation), the reaction of the fluorochromes to light is used to manipulate the volume of sample that actually fluoresces. Current Opinion in Biotechnology 2005, 16:3–12

MPM has become a powerful method for imaging thick samples, as it causes minimal harm to the surrounding cells or tissues through the unwanted absorption of short wavelength excitation energy [25]. This important advantage is achieved while still maintaining a good spatial resolution owing to the long wavelength used in the excitation (which is less damaging and penetrates deeper into tissues) and the efficient fluorescence detection that is spectrally far from the excitation. Typically excitation is in the infrared spectrum, whereas detection is in the visible spectrum. www.sciencedirect.com

From micro to nano: recent advances in high-resolution microscopy Garini, Vermolen and Young 9

Other non-linear methods include second harmonic generation (SHG) and coherent anti-Stokes Raman scattering (CARS). SHG was demonstrated to have advantages in thick samples and for the fast recording of small intensity changes, such as the detection of action potentials [26]. CARS provides the advantage that no labels are required, because the Raman-based method is sensitive to the specific chemical in the tissue itself [27,28]. Stimulated emission depletion

Stimulated emission depletion (STED) microscopy was conceptually introduced a decade ago by Hell and colleagues [29] and has been demonstrated recently [30]. The principle of the method is to ensure that the volume that emits fluorescence in the sample is extremely small. This is accomplished by using two pulsed lasers. The first laser has a wavelength that excites the fluorescent molecules and the second illuminates the sample with a donutlike pattern in a wavelength that drives (depletes) the excited states of the fluorescent molecules back to the ground state. The only volume that is left with fluorescent molecules in the excited state is in the hole of the donut, from where fluorescence is actually detected. Because pulsed lasers are used, the depleting laser is very powerful and the only molecules that are not fully depleted are very close to the donut hole. Images with a resolution of
30 nm have been measured using this technique [31], and given that the method is still in its infancy its potential is very high. This is a major step towards improving resolution: going from
100 nm (with 4Pi, I5M and HELM) to
30 nm (Figure 7) [32].

Saturated structured illumination microscopy

Saturated structured illumination (SSIM) is conceptually the opposite of STED. The theory was established by Heintzmann and colleagues [33,34] and was recently successfully tested by Gustafsson and colleagues [35]. By using a structured light illumination from two powerful interfering beams, most of the fluorescent molecules in the illuminated beams saturate, leaving only small volumes unsaturated at the shadows of the interference pattern. These volumes become very small when increasing the light intensity, smaller than any PSF width. The sample has to be scanned and the images have to be processed to extract the high-resolution data [35]. Resolutions smaller than 50 nm have already been demonstrated using this method (Figure 8).

High-resolution measurements of surfaces A few methods are suitable for high-resolution measurements of surfaces, but not for three-dimensional measurements. These include near-field scanning optical microscopy (NSOM) and methods that create entangled fields on the surface, including surface plasmon resonance (SPR) and total internal reflection fluorescence (TIRF). Near-field scanning optical microscopy

NSOM improves optical resolution by circumventing the use of lenses [36]. This is achieved by shining laser light through a fiber with a small tip aperture (diameter of 20– 200 nm). The tip scans the surface at a close distance (
10–50 nm) and the effective illuminated area has the tip diameter with a shallow penetration depth (less than

Figure 7

Stimulated emission depletion (STED) microscopy. (a) Confocal microscope image compared with (b) STED-4Pi microscopy of a fluorescently labeled microtubule of a human embryonic kidney labeled with a red-emitting dye (MR 121SE). The images are shown in the xz plane (‘side view’). The line on the left-hand side of the STED image is the result of a fluorescent layer attached to the microscope cover-slip. Pixel size is 95  9.8 nm along x and z, respectively [32]. (Figure reproduced from [32] with kind permission of SW Hell). www.sciencedirect.com

Current Opinion in Biotechnology 2005, 16:3–12

10 Analytical biotechnology

Figure 8

Improvement of resolution by structured illumination microscopy and nonlinear saturated structured illumination microscopy (SSIM). (a) Conventional microscope image, (b) structured illumination microscopy and (c) SSIM. The sample consists of fluorescent polystyrene beads with a nominal diameter of 51 nm [35]. (Figure reproduced with kind permission of M Gustafsson).

100 nm). Developed more than two decades ago [37,38], NSOM has gone through major improvements such as fast measurement (100 images per second) [39] and the detection of single molecules on surfaces, such as the membrane of dendritic cells [40]. Total internal reflection fluorescence and surface plasmon resonance

TIRF and SPR provide intrinsic high axial resolution owing to the unique optical setup that is used to illuminate the sample. A sample is attached to a substrate such as a glass cover slip adjacent to a prism that is illuminated at an angle that does not allow the light to propagate directly through the substrate to the sample. It is, however, enough to create an evanescent light field in a thin layer close to the substrate (
100 nm) and can excite fluorescent molecules that are within this shallow layer [41] (see also the review by Schneckenburger in this issue). The fluorescence emission is collected with an objective lens and thus provides a high axial resolution (
100 nm or less). TIRF has been recently applied to the dynamic study of molecular motor steps by combining it with an interference pattern illumination; a spatial resolution of 8 nm has been achieved with a time resolution of 100 ms [42]. In SPR, an evanescent field is also created, but here a glass surface is coated with a metal layer. Subtle changes in the refractive index of the surface can be detected by measuring changes in the intensity of the illumination beam as a function of the impinging angle [43]. This method is highly effective for detecting interactions at the surface without the need for a label. An attempt is currently underway to extend the method so that full imaging on a surface can also be achieved [44]. We believe that this method will have a number of advantages for monitoring the physical properties of biomacromolecules and protein–protein interactions. It Current Opinion in Biotechnology 2005, 16:3–12

has also been shown that SPR can be combined with nanostructured devices to provide an extraordinary behavior of light when it is transmitted through nanometersize holes [45]. This has the potential to evolve into a new type of microscopy with a resolution that is yet to be explored [46].

Conclusions For many years, it was believed that optical microscopy measurements had to be limited to a spatial resolution of
200 nm in the plane and
400 nm along the optical axis. Major developments, especially in the past few years, have demonstrated that this can be improved up to tenfold. Among the novel methods that have enabled these improvements, we believe that the non-linear methods combined with high numerical aperture measurements (e.g. using STED and SSIM) have the potential to evolve into systems that will be widely used, although the process may still take years. Furthermore, sophisticated methods are still being developed to improve the optical resolution for situations that involve thick samples (SPIM), fast measurements (SHG) and living cells (MPM). These methods will significantly improve the efficiency of research on biological samples and will provide new and necessary tools for subcellular and molecular level studies.

Acknowledgements We would like to thank Stefan W Hell (MPI Go¨ ttingen), Mats Gustafsson (UCSF), Rainer Heintzmann (MPI Go¨ttingen), Ernst HK Stelzer (EMBL) and Tziki Kam (Weizmann Institute) for providing helpful information and data to the review. Further we would like to acknowledge the Physics for Technology program of the Foundation for Fundamental Research in Matter (FOM, The Netherlands) for its support. www.sciencedirect.com

From micro to nano: recent advances in high-resolution microscopy Garini, Vermolen and Young 11

References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as:  of special interest  of outstanding interest 1.

Abbe E: Beitrage zur Theorie des Mikroskops und der Mikroskop ischen Wahrnehmung. Arch Mikroskop Anat 1873, 9:413-468. [Translation: Contributions to the theory of the microscope and that microscopic perception].

2.

Hooke R: Micrographia. London: Royal Society of London; 1664.

3.

Kam Z, Hanser B, Gustafsson MGL, Agard DA, Sedat JW: Computational adaptive optics for live three-dimensional biological imaging. Proc Natl Acad Sci USA 2001, 98:3790-3795.

4.

Booth MJ, Neil MAA, Jusˇ kaitis R, Wilson T: Adaptive aberration correction in a confocal microscope. Proc Natl Acad Sci USA 2002, 99:5788-5792.

5.

Schaefer LH, Schuster D, Herz H: Generalized approach for accelerated maximum likelihood based image restoration applied to three-dimensional fluorescence microscopy. J Microsc 2001, 204:99-107.

6.

Chuang TCY, Moshir S, Garini Y, Chuang A.Y-C, Young IT, Vermolen B, Doel Rvd, Mougey V, Perrin M, Braun M, et al.: The three-dimensional organization of telomeres in the nucleus of mammalian cells. BMC Biol 2004, 2:12:1-8.

7.

Minsky M: Microscopy apparatus. US Patent 3013467, 1961.

8.

Wilson T: Confocal Microscopy. London: Academic Press; 1990.

9.

Stephens DJ, Allan VJ: Light microscopy techniques for live cell imaging. Science 2003, 300:82-86.

10. Huisken J, Swoger J, Bene FD, Wittbrodt J, Stelzer EHK: Optical sectioning deep inside live embryos by selective plane illumination microscopy. Science 2004, 305:1007-1009. 11. Bailey B, Farkas DL, Taylor DL, Lanni F: Enhancement of axial resolution in fluorescence microscopy by standing-wave excitation. Nature 1993, 366:44-48. 12. Lanni F, Bailey B, Farkas DL, Taylor DL: Excitation field synthesis as a means for obtaining enhanced axial resolution in fluorescence microscopes. Bioimaging 1993, 1:187-196. 13. Wells WA: Man the nanoscopes. J Cell Biol 2004, 164:337-340. 14. Neil MAA, Jusˇ kaitis R, Wilson T: Method of obtaining optical sectioning by using structured light in a conventional microscope. Opt Lett 1997, 22:1905-1907. 15. Hell SW, Stelzer EHK: Fundamental improvement of resolution with a 4Pi-confocal fluorescence microscope using twophoton excitation. Opt Commun 1992, 93:277-282. 16. Schrader M, Hell SW, van der Voort HTM: Three-dimensional super-resolution with a 4Pi-confocal microscope using image restoration. J Appl Phys 1998, 84:4033-4042. 17. Egner A, Jakobs S, Hell SW: Fast 100-nm resolution three dimensional microscope reveals structural plasticity of mitochondria in live yeast. Proc Natl Acad Sci USA 2002, 99:3370-3375. Observation of mitochondria in living yeast at high resolution. 18. Egner A, Verrier S, Goroshkov A, So¨ ling H-D, Hell SW: 4Pi microscopy of the Golgi apparatus in live mammalian cells. J Struct Biol 2004, 147:70-76. Observation with high-resolution (
100 nm) optical microscopy of Golgi apparatus in living mammalian cells. 19. Gustafsson MGL, Agard DA, Sedat JW: Sevenfold improvement of axial resolution in 3D widefield microscopy using two objective lenses. Proc SPIE 1995, 2412:147-156. 20. Gustafsson MG, Agard DA, Sedat JW: 3D widefield microscopy with two objective lenses: experimental verification of improved axial resolution. Proc SPIE 1996, 2655:62-66. www.sciencedirect.com

21. Gustafsson MGL, Agard DA, Sedat JW: I5M: 3D widefield light microscopy with better than 100 nm axial resolution. J Microsc 1999, 195:10-16. 22. Frohn JT, Knapp HF, Stemmer A: True optical resolution beyond the Rayleigh limit achieved by standing wave illumination. Proc Natl Acad Sci USA 2000, 97:7232-7236. 23. Frohn JT, Knapp HF, Stemmer A: Three-dimensional resolution enhancement in fluorescence microscopy by harmonic excitation. Opt Lett 2001, 26:828-830. 24. Denk W, Strickler JH, Webb WW: Two-photon laser scanning fluorescence microscopy. Science 1990, 248:73-76. 25. Zipfel WR, Williams RM, Webb WW: Nonlinear magic: multiphoton microscopy in the biosciences. Nat Biotechnol 2003, 21:1369-1377. 26. Dombeck DA, Blanchard-Desce M, Webb WW: Optical recording of action potentials with second-harmonic generation microscopy. J Neurosci 2004, 24:999-1003. 27. Dudovich N, Oron D, Silberberg Y: Single-pulse coherently controlled nonlinear Raman spectroscopy and microscopy. Nature 2002, 418:512-514. 28. Heinrich C, Bernet S, Ritsch-Marte M: Wide-field coherent antiStokes Raman scattering microscopy. Appl Phys Let 2004, 84:816-818. 29. Hell SW, Wichmann J: Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy. Opt Lett 1994, 19:780-782. 30. Westphal V, Kastrup l, Hell SW: Lateral resolution of 28 nm (l/25)  in far-field fluorescence microscopy. Appl Phys B 2003, 77:377-380. Demonstration of axial resolution of
30 nm with stimulated depletion microscopy. 31. Hell SW: Toward fluorescence nanoscopy. Nat Biotechnol 2003, 21:1347-1355. 32. Dyba M, Jakobs S, Hell SW: Immunofluorescence stimulated  emission depletion microscopy. Nat Biotechnol 2003, 21:1303-1304. Demonstration of axial resolution of
30 nm with stimulated depletion microscopy on immunofluorescently labeled samples. 33. Heintzmann R, Jovin TM, Cremer C: Saturated patterned  excitation microscopy — a concept for optical resolution improvement. J Opt Soc Am A Opt Image Sci Vis 2002, 19:1599-1609. Theoretical description of SSIM that provides high resolution. 34. Heintzmann R: Saturated patterned excitation microscopy with two-dimensional excitation patterns. Micron 2003, 34:283-291. 35. Gustafsson MGL, Shao L, Agard DA, Sedat JW: Nonlinear structured illumination microscopy. In Focus on Microscopy 2003 April 13-16 2003; University of Genova, Italy: 2003, Abstract number 51. 36. Lewis A, Taha H, Strinkovski A, Manevitch A, Khatchatouriants A, Dekhter R, Ammann E: Near-field optics: from subwavelength illumination to nanometric shadowing. Nat Biotechnol 2003, 21:1378-1386. 37. Lewis A, Isaacson M, Harootunian A, Muray A: Development of a 500 A˚ spatial resolution light microscope I. Light is efficiently transmitted through l/16 diameter apertures. Ultramicroscopy 1984, 13:227-231. 38. Betzig E, Trautman JK, Harris TD, Weiner JS, Kostelak RL: Breaking the diffraction barrier: optical microscopy on a nanometric scale. Science 1991, 251:1468-1470. 39. Humphris ADL, Hobbs JK, Miles MJ: Ultrahigh-speed scanning  near-field optical microscopy capable of over 100 frames per second. Appl Phys Let 2003, 83:6-8. High-speed measurements with a near-field microscope. 40. Koopman M, Cambi A, de Bakker BI, Joosten B, Figdor CG, van Hulst NF, Garcia-Parajo MF: Near-field scanning optical microscopy in liquid for high resolution single molecule detection on dendritic cells. FEBS Lett 2004, 573:6-10. Current Opinion in Biotechnology 2005, 16:3–12

12 Analytical biotechnology

41. Axelrod D: Total internal reflection fluorescence microscopy in cell biology. Traffic 2001, 2:764-774.

44. Stabler G, Somekh MG, See CW: High-resolution wide-field surface plasmon microscopy. J Microsc 2004, 214:328-333.

42. Cappello G, Badoual M, Ott A, Prost J, Busoni L: Kinesin motion in the absence of external forces characterized by interference total internal reflection microscopy. Phys Rev E 2003, 68, 021907:1-7.

45. Barnes WL, Dereux A, Ebbesen TW: Surface plasmon subwavelength optics. Nature 2003, 424:824-830.

43. Notcovich AG, Zhuk V, Lipson SG: Surface plasmon resonance phase imaging. Appl Phys Let 2000, 76:1665-1667.

Current Opinion in Biotechnology 2005, 16:3–12

46. Garini Y, Kutchoukov VG, Bossche A, Alkemade PFA, Docter M, Verbeek PW, Vliet LJv, Young IT: Toward the development of a three-dimensional mid-field microscope. Proc SPIE 2004, 5327:115-122.

www.sciencedirect.com