UV-laser-based microscopic dissection of tree rings – a novel sampling tool for δ13C and δ18O studies

Authors


Summary

  • UV-laser-based microscopic systems were utilized to dissect and sample organic tissue for stable isotope measurements from thin wood cross-sections.
  • We tested UV-laser-based microscopic tissue dissection in practice for high-resolution isotopic analyses (δ13C/δ18O) on thin cross-sections from different tree species. The method allows serial isolation of tissue of any shape and from millimetre down to micrometre scales. On-screen pre-defined areas of interest were automatically dissected and collected for mass spectrometric analysis.
  • Three examples of high-resolution isotopic analyses revealed that: in comparison to δ13C of xylem cells, woody ray parenchyma of deciduous trees have the same year-to-year variability, but reveal offsets that are opposite in sign depending on whether wholewood or cellulose is considered; high-resolution tree-ring δ18O profiles of Indonesian teak reflect monsoonal rainfall patterns and are sensitive to rainfall extremes caused by ENSO; and seasonal moisture signals in intra-tree-ring δ18O of white pine are weighted by nonlinear intra-annual growth dynamics.
  • The applications demonstrate that the use of UV-laser-based microscopic dissection allows for sampling plant tissue at ultrahigh resolution and unprecedented precision. This new technique facilitates sampling for stable isotope analysis of anatomical plant traits like combined tree eco-physiological, wood anatomical and dendroclimatological studies.

Introduction

Stable isotope analyses (hydrogen (H), nitrogen (N), carbon (C), oxygen (O)) on organic tissue are widely used in plant physiological, ecological and climatological research (Ehleringer et al., 1993; Ziegler, 1995; Schleser et al., 1999; Dawson et al., 2002; Helle & Schleser, 2004b; Treydte et al., 2006; West et al., 2010; Hietz et al., 2011; Brienen et al., 2012; Werner et al., 2012; Heinrich et al., 2013). State-of-the-art isotope ratio mass spectrometry (IRMS) and cavity ring-down spectroscopy (CRDS) allow precise determination of stable isotope ratios on minimal sample amounts and large numbers of samples. However, accurate preparation of small samples in large numbers is challenging, as this still has to be done manually in most of the research areas. Despite the high precision and accuracy of modern analytical devices, data quality very much depends on the skills and motivation of the person responsible for sample preparation. This applies in particular to inter- and intra-annual stable isotope studies of woody tissue attempting to extract valuable seasonal climatic and environmental information (Leavitt & Long, 1982; Loader et al., 1995; Treydte et al., 2006; Schollaen et al., 2013) or assessing plant eco-physiological processes (Helle & Schleser, 2004a; Gessler et al., 2009; Schubert & Jahren, 2011; Krepkowski et al., 2013) from annual growth rings or parts thereof. Furthermore, high-precision dissection is crucial in studies using high-resolution stable isotope data for the identification of anatomically nondistinct annual rings in tropical trees (Evans & Schrag, 2004; Verheyden et al., 2004; Pons & Helle, 2011).

Different methods exist for the dissection of wood tissue from tree rings for stable isotope measurements. One common method has been to divide wood segments or cores by cutting tangential slices, utilizing fixed-blade sledge and rotary microtomes (Ogle & McCormac, 1994; Loader et al., 1995; Barbour et al., 2002; Helle & Schleser, 2004a; Poussart et al., 2004; Verheyden et al., 2004; Ogée et al., 2009; Pons & Helle, 2011) or scalpel/razor blades (Roden et al., 2009; Managave et al., 2010). With a microtome, wood slices down to c. 10 μm thickness can be achieved (Helle & Schleser, 2004a). Another high-resolution sampling method is by collecting wood dust from a series of small wood holes drilled in the radial direction with twist drills (Walcroft et al., 1997; Gebrekirstos et al., 2009; Fichtler et al., 2010), robotic micromilling techniques (Wurster et al., 1999; Dodd et al., 2008) or an UV-laser ablation system in combination with isotope ratio mass spectrometry (LA-C-GC-IRMS), first described in Schulze et al. (2004). Accurate sample adjustment, as well as unambiguous identification of tree rings, is normally provided by visual inspection using a microscope or digital camera. After dissection, wood samples are ground to a fine powder and either cellulose is extracted or bulk wood samples are used for stable isotope analysis. The latter is the case for the UV-laser ablation method where dust is ablated from a wood core or segment in an airtight chamber flushed with helium. From the ablation chamber wood dust is directly transferred into a combustion oven and converted to CO2. Subsequently, δ13C is measured by an isotope ratio mass spectrometer that is coupled online to the UV-laser ablation and preparation system. Various studies used this method to analyse the intra-annual δ13C in tree rings responding to carbohydrate storage and remobilization, as well as short-term climatic effects in conjunction with wood density variability (Intra-annual Density Fluctuations, IADFs) (Skomarkova et al., 2006; Battipaglia et al., 2010; De Micco et al., 2012). The spatial resolution of a UV-laser shot is 40 μm (Schulze et al., 2004) to 150 μm (De Micco et al., 2012). The widths of the holes produced by the mechanical drills can vary from 500 μm (Gebrekirstos et al., 2009) to 800 μm (Fichtler et al., 2010) up to 1.5 mm (Walcroft et al., 1997) depending on the diameter of the drill. Robotic micromilling generally requires a minimum path width of 400 μm for increment core samples, whereas individual path widths of < 100 μm are possible (Dodd et al., 2008).

Different drawbacks exist for the dissection methods mentioned before. For the methods using blades or microtomes, wood samples must have straight-line ring structure boundaries in both tangential and transversal directions, as samples are cut in parallel linear segments. However, ring width and curvature of growth rings are rarely strictly parallel and fibre or vessel angles vary both under natural conditions and depending on species-specific wood anatomical characteristics. Thus, often only a limited number of consecutive tree rings per sample is suitable for high-precision intra-annual stable isotope measurements when applying these dissection methods. Generally, cross-contamination during cutting, milling and other preparatory steps cannot be excluded especially when quality control by electronic image documentation is not implemented. For the methods involving drilling wood holes, straight-line boundaries in the transversal direction are less important, but differences in vertical direction can be problematic, depending on the drill depth and the deviation of wood fibre or vessel orientation from the vertical (Fig. 1b). LA-C-GC-IRMS systems are presently confined to carbon isotope analysis. Aerosol transfer of wood dust between the ablation chamber and the combustion furnace requires a rather high helium consumption and transfer capillaries sometimes may clog up.

Figure 1.

Possible steps of wood sample preparation and high-precision sampling using UV-laser-based microscopic dissection systems. (a) Top view and profile of a metal frame slide with wood cross-sections used for gravity-assisted collection (LMD7000, Leica). Up to three cross-sections of 59-mm length can be mounted between two metal frames. For stable isotope analyses it is possible to cut serial sections at inter- to intra-annual resolution or even to pool sample material, for example, for cellulose extraction. To avoid material loss the dissection of tissue can be alternatingly shifted. (b) Schematic view of sample preparation with a wood segment in radial and cross-section. The use of cross-sections diminishes difficulties of cross-contamination when sampling tree rings with tight curvatures, nonparallel boundaries, varying fibre and vessel angles, and/or changing tree-ring width in the radial direction. The dissection of wood cross-sections ranging in thickness from 100 μm to a maximal 1000 μm is possible. The remaining wood sample can be used for multidisciplinary approaches, for example, combining high-resolution stable isotope analysis, quantitative wood anatomy, ring width or wood density analysis.

The need for methodological improvements to facilitate fine dissection of irregular and narrow shapes has emerged lately, so we assessed a novel method to dissect wood tissue from tree rings using the UV-laser microdissection microscopes produce by Leica (LMD7000; Leica Microsystems GmbH, Wetzlar, Germany) and Zeiss (PALM MicroBeam; Carl Zeiss Microscopy GmbH, Jena, Germany). UV-laser-based microscopic dissection systems are being applied widely in biomedical (Fend & Raffeld, 2000), animal (Bayona-Bafaluy et al., 2005) and also in plant research (Nelson et al., 2006; Abbott et al., 2010). However, to our knowledge, this is the first time it has been utilized to precisely dissect plant tissue for stable isotope research. Laser-based microscopic dissection uses an UV-laser beam to isolate tissues of interest from thin sections of samples. We describe the design and handling of the two different UV-laser microdissection microscopes. Advantages and constraints are discussed on the basis of high-resolution stable isotope analyses on woody plant material of deciduous and coniferous trees from temperate and tropical climate zones. Variability of the new sub-seasonal C and O stable isotope data was evaluated with respect to seasonal changes of climatic conditions and wood growth dynamics. Furthermore, we tested whether δ13C in the woody ray parenchyma and in xylem cells of a deciduous tree is different.

Materials and Methods

UV-laser-based microscopic dissection

UV-laser microdissection microscopes were used as a precise tool for dissecting different wood samples in stable isotope studies. We tested two different UV-laser microdissection microscopes of Leica (LMD7000, software v6.7.1.3952, Leica Microsystems GmbH) and Zeiss (PALM MicroBeam, software v4.5.0.9, Carl Zeiss Microscopy GmbH). We followed a preparation scheme containing five steps:

  1. manual preparation of thin wood cross-sections (max. 1000 μm thickness) with a microtome or a high-precision saw,
  2. microscopic identification and pen-screen selection of tissues of interest,
  3. automatic UV-laser-based microscopic dissection of inter-or intra-annual wood sections,
  4. semi-automatic sample collection into tin or silver capsules by gravity or forceps, optional chemical treatment (e.g. cellulose extraction), and
  5. stable carbon (C) and oxygen (O) isotope analysis via conventional Isotope Ratio Mass Spectrometry (IRMS) coupled online to a combustion or pyrolysis furnace.

UV-laser-based microscopic dissection enables selection of relevant plant cells/tissues on screen by pen, while nonrelevant tissues (e.g. resin ducts, wood rays) may not be selected or be removed (Fig. 2a,d) before sampling. Any size and area can be dissected, which is important for the precise dissection of asymmetric tree rings or parts thereof, as shown for example by lobate growth (Fig. 2a), intra-ring density fluctuations (Fig. 2b) or wedging tree rings (Fig. 2c). Furthermore, it is possible to cut serial sections or even to pool sample material, for example if the weight of the dissected tissue from one thin section is insufficient for a stable isotope measurement. Sample material from the same array of wood cells can be identified unambiguously on a second or third cross-section and may be pooled for chemical treatment, like cellulose extraction before IRMS analysis (Fig. 1a).

Figure 2.

Schematic examples of different dissected wood tissues from several tree species for cutting process with an UV-laser microscopic dissection system. Nonrelevant tissues (e.g. resin ducts, wood rays) can be removed before stable isotope analysis and any size and area can be dissected. (a) Cross-section of Pinus sylvestris. Resin ducts and asymmetric growth rings can be easily dissected. (b) Cross-section of P. sylvestris. Density fluctuations within tree rings are dissected (nos. 2 and 5). (c) Cross-section of Juniperus excelsa. Tree rings are marked numerically, including wedging rings. (d) Cross-section of Quercus robur. Wood ray (no. 1) is marked together with tree rings.

In general, wood samples of 100-μm to 1000-μm thickness can be dissected without any major constraints (Fig. 1b). Dissection samples of up to 1000-μm thickness are achieved by cutting iterations and adjustment of the z-focus of the UV-laser beam. The use of cross-sections thinner than 100 μm normally results in insufficient amounts of material for present-day combustion/pyrolysis systems normally coupled to IRMS. Taking thin sections, however, has the advantage that part of the wood sample is left intact, providing the opportunity for reanalysis of ring widths or allowing other investigations, such as, for example, quantitative wood anatomy or wood density measurements. Furthermore, the risk of cross-contamination of various tissues is diminished, for example when studying narrow tree rings that have nonparallel boundaries and/or have varying fibre and vessel angles (Fig. 1b).

Sample material

In order to test the UV-laser-based microscopic dissection method in practice we focused on three applications. In the first experiment, we tested whether the δ13C values in xylem cells of an oak sample (Quercus robur L.) differ from the δ13C values in ray cells in order to assess potential influences of ray parenchyma on the isotope variability of tree-ring sequences. Therefore, an oak tree was sampled at Telegrafenberg Hill in Potsdam, northeastern Germany (52°23′N, 13°04′E; 94 m above sea level (a.s.l.)). Germany is characterized by a warm, temperate and humid climate of the mid-latitudes with weakening oceanic influences from the northwest to the southeast. In the second and third example, we dissected wood tissues to analyse high-resolution intra-annual δ18O, δ13C and wood density data. The material selected comprises wood from coniferous and broadleaved trees from two study sites differing in climatic conditions. The sampling site of the coniferous tree (Pinus strobus L.) is at St Arnold, northwestern Germany (52°13′N, 7°23′E; 56 m a.s.l.), having similar climatic conditions as the site of the first example. The sampling site of the broadleaved tree is in lowland rainforest in the eastern part of Central Java, Indonesia (07°52′S, 111°11′E; 380 m a.s.l.). A 5-mm diameter increment core of a living teak tree (Tectona grandis L.) was chosen from a collection that was gathered during a field campaign in 2008 (Schollaen et al., 2013). At this study site, the climate is characterized by a distinct dry season from June to September and a rainy season from October to May (Fig. 5b). The growing season for teak in this region generally lasts from the beginning of October to the end of May. During the dry season the trees are leafless and produce no wood as they are in a state of cambial dormancy (Coster, 1928).

Sample preparation

First, increment cores were cut into segments of 5 cm length and second, transverse or cross-sections of c. 500-μm thickness for the teak sample, 350-μm thickness for the oak sample and 1000-μm thickness for the pine sample were cut with a high-precision diamond saw (ISOMED5000, ITW Test & Measurement GmbH, Düsseldorf, Germany) or a core microtome (Gärtner & Nievergelt, 2010), respectively. Extractives, such as resins and oils, were removed from the wood cross-sections by boiling them in de-ionized water and ethanol. For further treatment, the cross-sections of resin-extracted wood were fixed between two stainless metal frame slides (Fig. 1a) for use with the Leica system and between two conventional microscope slides (26 × 76 mm, Thermo Fisher Scientific; Menzel GmbH, Braunschweig, Germany) for the Zeiss system. Three of these slides, each carrying up to three cross-sections of maximum 6-cm length, can be mounted onto the slide holder of each microscopic dissection system. Hence, in total cross-sections from a wood core measuring c. 54 cm in length can be processed by a UV-laser-based dissection system in one operation.

Sample collection

Depending on the thickness of the wood cross-sections, samples are viewed under the microscopes in transmitted- or reflected-light mode. Series of tissues of interest are first marked with mouse or screen-pen (Figs 2, 3). Every segment drawn was dissected with the UV-laser beam and collected for the Leica system by gravity into single silver (δ18O) or tin (δ13C) capsules standing in a collection holder. For the Zeiss system tissue was taken up with a forceps and put into silver/tin capsules.

Figure 3.

Microscopic image of a teak (Tectona grandis) cross-section (500 μm thickness, magnification ×5). Consecutive sections of intra-annual tree-ring parenchyma and xylem tissue marked with numbers for the cutting process with an UV-laser microdissection microscope (LMD7000; Leica). Screen shots before, during and after UV-laser-based microscopic dissection of wood parenchyma band.

In the first experiment on an oak sample (Q. robur), resin-extracted wood and cellulose of ray parenchyma and xylem tissue from six (cellulose) and eight (wood) consecutive tree rings were cut, respectively (Fig. 4a). Cellulose extraction was performed according to the standard method as described by Wieloch et al. (2011).

Figure 4.

Comparison between δ13C values from resin-extracted wood (δ13Cwd) and cellulose (δ13Ccel) of ray parenchyma and xylem tissue in tree rings of an oak sample (Quercus robur). (a) Microscopic image of an oak cross-section (350-μm thickness) with schematic example of dissected tree-ring xylem and ray tissues for cutting process with the UV-laser-based microscopic dissection method. (b) Tree-ring δ13C values (δ13Cwd) of ray parenchyma and xylem cells from six consecutive tree rings dissected with the UV-laser microscope. (c) Cellulose δ13C values (δ13Ccel) of ray parenchyma and xylem cells from eight consecutive tree rings. The analytical precisions of ± 0.08‰ are shown as error bars. The δ13C signals of ray and xylem tissue follow the same year-to-year variability and trend, but reveal offsets that are opposite in sign depending on whether wholewood or cellulose is analysed.

In the second and third experiment, tree-ring parts of interest were graphically subdivided on a pen-screen in the radial direction into equidistant parts of 150-μm width for the teak sample (Fig. 3) and 20-μm width for the pine sample. Cutting lines were drawn parallel to wood anatomical structures and tree-ring boundaries independent of their shape. The number of segments per year varied depending on the tree-ring width.

Stable isotope analyses

Oxygen isotope ratios were measured using a high temperature TC/EA pyrolysis oven (at 1340°C) coupled online to an Isotope Ratio Mass Spectrometer (IRMS; Delta V Advantage; Thermo Fisher Scientific, Bremen, Germany). The carbon isotope ratios were measured by combustion (at 1080°C) using an elemental analyser (Model NA 1500; Carlo Erba, Milan, Italy) coupled online to an IRMS (Isoprime Ltd, Cheadle Hulme, UK). Sample masses of 130–220 μg of resin-extracted wood or cellulose were used for IRMS analyses. Sample replication resulted in a reproducibility of better than ± 0.1‰ for δ13C values and up to ± 0.3‰ for δ18O values. The isotope ratios are given in the delta (δ) notation, relative to the standards VPDB for δ13C and VSMOW for δ18O (Craig, 1957).

Determination of wood density

Wood density of the intra-annually resolved pine sample was precisely calculated from the masses of equidistant subsections of equal volume (20 × 1000 × 4000 μm) that were weighed by a micro balance (AX26 DeltaRange; Mettler Toledo GmbH, Greifensee, Switzerland).

Results

Influence of ray parenchyma on the δ13C signature in tree rings of oak (Quercus robur)

All δ13C values from resin-extracted wood (δ13Cwd; Fig. 4b) of the ray parenchyma were found to be lower than those of the associated xylem tissue, with a mean difference of 0.21‰. This is well above the analytical precision of ± 0.08‰. Despite the significant general offset, δ13C signals of ray and xylem tissue follow the same year-to-year variability and trend as confirmed by a R² of 0.99. Likewise, cellulose δ13C values (δ13Ccel; Fig. 4c) from ray and xylem tissue showed the same year-to-year variability. However, the δ13Ccel values of ray parenchyma were slightly higher compared to the δ13Ccel values of xylem tissue. The mean difference is 0.26‰ with an R² of 0.96 and analytical precision of ± 0.08‰.

High-resolution intra-annual δ18O results from teak (Tectona grandis)

The intra-annual δ18O profiles of a teak tree (δ18Owd) from Indonesia showed a clear annual cycle (Fig. 5a). Annual wood formation starts with a parenchyma band showing δ18Owd values that are similar to the δ18Owd values at the end of the previous tree ring. Wood formed directly after the parenchyma band is characterized by rapidly rising δ18Owd values up to the seasonal maximum which appears early in the growing season. This δ18Owd maximum is followed by a decline to a seasonal minimum typically in the 2nd third of each tree ring before δ18Owd marginally rises again in the last third of the growing season. The pattern described is rather consistent in spite of the different numbers of sub-sections per year and follows the annual cycle in rainfall amount and its corresponding isotope signature (Fig. 5b). The high-resolution pattern for 1985 does not show distinct high values at the beginning of the growing season and, as in 1984, reveals two minima during the growing season. Note that for tree-ring dating we followed the convention for the southern hemisphere, which assigns to each tree ring the year in which radial growth begins (Schulman, 1956).

Figure 5.

Intra-annual δ18Owd variations from teak (Tectona grandis) collected in the eastern part of Central Java, Indonesia and corresponding precipitation data. (a) High-resolution isotope data are plotted vs monthly precipitation sums at the study site (GPCC V6, mean of 9 nearest grid-points, Schneider et al., 2011) and δ18O values in precipitation (δ18OPre) of the corresponding year from Jakarta/Indonesia (IAEA/WMO, 2006). Dissected part of parenchyma band is marked as a dashed line. El Niño and La Niña events are highlighted in yellow and blue, respectively (NOAA, 2013). (b) Average monthly precipitation sums (bars) for the eastern part of Central Java (GPCC V6, 1901–2010, mean of 9 nearest grid-points, Schneider et al., 2011) and mean δ18O in precipitation (line) for Jakarta/Indonesia (1962–1998) (IAEA/WMO, 2006) are shown. The seasonal δ18Owd pattern reflect the annual cycle of δ18O in precipitation and corresponding seasonal changes in the amount of precipitation at the study site (Schollaen et al., 2013). δ18Owd pattern is sensitive to rainfall extremes caused by ENSO events.

High-resolution intra-annual δ13C, δ18O and wood density of white pine (Pinus strobus)

The δ13Cwd profile begins with a downward trend of c. 1‰ during the first third of the tree ring (Fig. 6b). δ13Cwd values rise again in the middle part before they increase strongly by > 2‰ to a seasonal maximum followed by a sharp and drastic decrease at the very end of the latewood. By contrast, the δ18Owd profile is characterized by a broad peak in the first third of the tree ring, reaching δ18Owd values of > 30‰. A second but less pronounced increase is indicated for the last part of the ring, with a maximum that is synchronous to the peak in δ13Cwd. The wood density profile starts with low values and follows synchronously, although attenuated, the increase and decline of δ18Owd within the first half of the tree ring. Starting from a minimum in the middle of the tree ring, wood density increases towards the end of the tree ring. Beside the general increase in density, a few intra-annual density fluctuations (IADFs) are visible. Increasing δ13Cwd values are related to these short-term intra-ring density fluctuations with a short time lag. At the end of the growing season all parameters (δ13Cwd, δ18Owd, wood density) reach their seasonal maxima.

Figure 6.

Seasonal variations of wood density and stable isotope ratios (δ13Cwd, δ18Owd) of a pine tree (Pinus strobus) from northwestern Germany for 1977 (b) and the corresponding daily temperature, precipitation and relative humidity data (a). Intra-annual density fluctuations (IADFs) are marked with arrows in the photomicrograph. Seasonal moisture signals in δ18Owd are weighted by nonlinear intra-annual growth dynamics. DOY, day of year.

The corresponding seasonal courses of temperature, precipitation and relative air humidity for 1977 do not deviate strongly from the long-term mean, except for a conspicuous dry period of 3 wk in May (Fig. 6a). Only one strong rainfall event was recorded during this period and relative humidity was persistently below 60%.

Discussion

Ray parenchyma and xylem tissue of deciduous trees show offsets in δ13C, but same year-to-year variability

Our results exhibit the same year-to-year variability between the δ13C signals of ray parenchyma and xylem tissue over 6 and 8 yr, respectively (Fig. 4). The similarity of data from the different woody tissues may indicate that they result from the same physiological and biochemical processes. Of course, during the vegetation period assimilates are used at the same time for vessel-, fibre- and ray parenchyma-formation. However, ray cells live longer, up to several years and thus their structure and chemical composition may change over longer time intervals. Furthermore, woody rays are the main storage organs for nonstructural carbohydrates, for example sugars and starch. Accumulation and mobilization of nonstructural carbohydrates may be accompanied by increased metabolic activity and respiration of parenchymatic tissue as compared to xylem tissue. Indeed, the offsets found between δ13C of ray parenchyma and xylem tissue seem to reflect differences in chemical composition and metabolic activity. Ray parenchyma dissected from resin-extracted wood is depleted in 13C as compared to xylem tissue (Fig. 4b). Our finding confirms the findings in a study by Vaganov et al. (2009) using LA-C-GC-IRMS on wood tissue of four consecutive tree rings from a Norway spruce (Picea abies) sample. Data revealed a similar 13C-depletion of ray parenchyma of up to 0.23‰ in comparison to tracheids. Vaganov et al. (2009) explained the lower δ13Cwd values as being a result of continuing incorporation of lignin into the long-lived parenchyma cells. Lignin is generally depleted in 13C by up to 3‰ as compared to cellulose (Wilson & Grinsted, 1977; Loader et al., 2003; Harlow et al., 2006). Hence, a slightly higher content of lignin in ray cells may well cause the observed 13C depletion over xylem tissue in coniferous wood (Vaganov et al., 2009) and in broadleaved oak of this study. However, UV-laser-based microscopic tissue dissection allows chemical treatment, like cellulose extraction, before IRMS analysis. In contrast to a depletion in 13C of resin-extracted wood we found cellulose of ray parenchyma (δ13Ccel) to be slightly enriched in 13C when compared to cellulose from xylem tissue. This minor, but apparent enrichment of cellulose of woody rays may be promoted by slightly enhanced catabolic metabolism, which is enhanced respiration making ATP available for the biochemical processes of accumulation and mobilization of starch. Gleixner et al. (1993) showed that ‘lighter’ sugar molecules are preferentially used in catabolic, or respiratory reactions, whereas ‘heavier’ ones are involved in polymerization of cellulose, for example. Hence, an increased carbon isotope partitioning between anabolic and catabolic metabolism in ray parenchyma may result in the observed slight enrichment in δ13CCel in comparison with xylem tissue. Note, that the differences in δ13C between woody rays and xylem tissue may not be constant (Fig. 6, Vaganov et al., 2009). Varying cellulose to lignin ratios or respiration rates may cause the observed differences in the offsets in δ13C and may also act on longer timescales.

A recent study on ray parenchyma from Spanish juniper (Juniperus thurifera) revealed that time series based on the abundance of ray cells reveal complementary climatic signals to those derived from tree-ring width chronologies (Olano et al., 2013). Our results demonstrate that the δ13C values of ray parenchyma and xylem tissue contain the same year-to-year variability. Hence, we can assume from this pilot study that the δ13C signals of ray cells may not provide additional information in dendroclimatic studies. However, the offsets in δ13C between ray cells and xylem are dependent on whether wood or cellulose is analysed. This result indicates the crucial value of the new dissection method as it allows for new approaches towards a better understanding of the physiological processes controlling ray parenchyma and xylem cell formation.

High-resolution δ18O values in tree rings of Indonesian teak are sensitive to rainfall extremes caused by ENSO

The seasonal tree-ring δ18Owd pattern in Indonesian teak is hypothesized to reflect the annual cycle of δ18O in precipitation and corresponding seasonal changes in the amount of precipitation at the site (Fig. 5). It has been shown that high δ18Owd values during the start of the growing season represent the δ18OPre signature of the prior dry season, while the lowest tree-ring δ18O values reveal the δ18OPre signature of the main rainy season (Schollaen et al., 2013). Towards the end of the growing season tree-ring δ18O values increase again following the δ18O-trend of precipitation. The described pattern fits closely during the years 1983 and 1986. The low rainfall amount during the dry season of 1983 was caused by an El Niño event (El Niño 1982–1983). The two minima in tree-ring δ18O values during the 1984, as well as the missing upward trend at the beginning of the growing season in 1985, can be explained by rather high rainfall amounts due to an ongoing La Niña phase (blue shaded periods). This demonstrates that sub-seasonal tree-ring δ18O records of Indonesian teak are very sensitive to rainfall extremes caused by ENSO, with high δ18Owd values during El Niño events and low δ18Owd values during La Niña events.

Correlations between δ18Owd and rainfall amount were also found in several other studies on tropical or subtropical trees (Poussart et al., 2004; Managave et al., 2010; Brienen et al., 2012; Sano et al., 2012). However, to our knowledge, this is the first time that intra-annual δ18Owd values in tropical trees have been shown to reflect the rainfall pattern over an entire year with distinct rainy and dry season signals. This underlines the value of high-resolution intra-seasonal isotope measurements of tropical wood, especially in the light of extreme rainfall events often associated with El Niño and La Niña.

High-resolution stable isotope variations in white pine reflect seasonal changes of climatic conditions weighted by nonlinear growth dynamics

Tree rings integrate environmental information over the vegetation period. The intra-annual variations in δ13Cwd, δ18Owd and wood density are driven mainly by interactions between seasonal variation in meteorological conditions, soil water availability and plant response. Specific weather events such as dry periods are clearly recorded in the seasonal pattern of δ13C, δ18O and wood density (Barbour et al., 2002; Eilmann et al., 2010; Sarris et al., 2013). In this study a detailed review of the weather conditions for 1977 at our site in northwestern Germany indicated several time periods with no or less rainfall – so called ‘dry’ days (Fig. 6a). The corresponding intra-annual pattern of δ13Cwd (Fig. 6b) follows a typical course observed for coniferous species with a gradual increase in δ13Cwd to a maximum in latewood followed by a sharp decline at the very end of the tree ring (cf. Leavitt, 1993; Walcroft et al., 1997; Barbour et al., 2002; Schulze et al., 2004). Short-term increases in δ13Cwd values of the second half of the tree ring were observed after dry periods and are in agreement with the wood density pattern. By contrast, the first dry period in May is not well observed in the δ13Cwd pattern. However, this dry period in late spring, characterized by low rainfall as well as low relative humidity, is well represented in the δ18Owd record by a conspicuous broad peak. It seems that δ18Owd is much more sensitive to changes in moisture conditions than δ13Cwd and wood density at the beginning of the growing season. During the second half of the growing season it appears to be vice versa, that is, δ13Cwd and wood density reflect moisture deficits better than δ18Owd does.

Furthermore, environmental information seems to be integrated in the tree ring in a nonlinear way. The ‘dry’ period of 21 d in May represents approximately only one-eighth of the vegetation period, while the δ18Owd peak represents one-third of the whole tree-ring width. The ability of a tree ring to integrate environmental information depends on the rate of wood cell formation and the longevity of cells. Within the vegetation period these two features are inversely related (Fig. 7). While thin-walled earlywood cells are formed at a high rate, they may only live for days up to a few weeks. Thick-walled latewood cells are built at very slow rates, but may live up to several months and integrate environmental information over a much longer period of time (Fig. 7a). Hence, environmental conditions reflected by intra-annual tree-ring parameters are weighted by seasonal dynamics of wood formation. Additionally, other seasonal changes related to tree physiological processes such as kind and size of carbohydrate pools (amount of C and O), partitioning between catabolic and anabolic metabolism, as well as metabolic flux rates in conjunction with corresponding isotope fractionations have to be considered (Fig. 7b and Helle & Schleser, 2004a; Gessler et al., 2009; Werner et al., 2012).

Figure 7.

(a) Nonlinear seasonal growth increments and longevity of cells during the vegetation period for conifers at temperate sites (DOY, day of year). (b) Seasonal dynamics of major factors influencing tree-ring stable isotope ratios of wood cell formation is characterized by increasing cell wall thickness, cell longevity and amount of C/O per cell row (elemental mass/wood density), and decreasing cell diameter. EW, early wood; LW, late wood.

Evaluation and comparison of UV-laser-based microscopic dissection systems

The results presented suggest that UV-laser-based microscopic dissection is a very useful method for sampling woody tissue in stable isotope studies. It can be used likewise for dissecting whole tree rings and intra-annual sampling at very high spatial resolution. The two UV-laser microdissection microscopes of Leica (LMD7000, software v6.7.1.3952, Leica Microsystems GmbH) and Zeiss (PALM MicroBeam, software v4.5.0.9, Carl Zeiss Microscopy GmbH), we tested in these studies, differ concerning their practical implementations and applications (Table 1). The laser from Leica is moved via optics and the cross-section samples are mounted on a fixed stage. The Leica system uses high-precision optics to steer the laser beam by means of prisms along the desired cut lines on the tissue. As a side effect, the laser can only cut drawn lines or areas marked within the actual microscopic field of view. If larger areas, for example whole tree rings, need to be dissected additional shapes must be defined and additional dissections are required for gathering a single large sample. The dissected sample tissues principally fall down by gravity into collection vessels, for example, tin or silver capsules. Thus, samples can be prepared directly for conventional autosampler systems coupled to isotope ratio mass spectrometers. An important limitation of the Leica system is the lack of an automatic z-focus adjustment that would allow repeated laser cutting of thicker cross-sections.

Table 1. Comparison of the characteristics of the two UV-laser-based microscopic dissection systems tested
Zeiss microdissection microscope (PALM MicroBeam/software v4.5.0.9)Leica microdissection microscope (LMD7000/software v6.7.1.3952)
General characteristics

Laser specifications:

 Wavelength: 355 nm

 Max. pulse energy: 100 μJ

 Laser beam stays fixed, stage is moved

Cutting process with different magnifications (e.g. ×2.5, ×5 or ×10)

Best range of sample thickness: 100–1000 μm

Laser specifications:

 Wavelength: 349 nm

 Max. pulse energy: 120 μJ

 Laser beam is moved, stage stays fixed

Cutting process with different magnifications (e.g. ×5, ×10 or ×20)

Best range of sample thickness: 100–1000 μm

Advantages

Selection of tissue/shape beyond the visible screen

Automatic z-focus adjustment/cutting iterations possible

Better software handling and multifarious choice of drawing tools

Laser has more power and less cutting iteration needed

Fast sample collection process by semi-automatic sample collection into tin or silver capsules by gravity

Automatic image documentation before and after each dissection by image-database (IM500)

Limitations

Nonautomatic collection of selected tissues of interests in tin/silver capsules (specimens are being picked up manually with a forceps and transferred into tin/silver capsules)

Laser has less power and frequent cutting iteration needed

Nonautomatic image documentation of the laser dissection process

Lack of automatic z-focus

Drawing area of interests larger than the field of view is not possible

Control check of dissected samples needed after every cutting cycle as samples may stick and do not fall down automatically

Compared to the Leica microscope, the objectives of the Zeiss microscope are installed inversely, or underneath the sample holder. Hence, tissues of interests are marked via mouse or screen pen on the lower sample side and can be selected beyond the visible screen. Thus, tissues of interests can be marked as big as necessary, because the laser stays fixed while the sample is moved by the high-precision stage during the dissection process. The UV-laser passes through the glass slide and the dissected sample tissues remain in position. After all of the marked tissue has been dissected specimens are picked up manually with a forceps and transferred into tin or silver capsules for stable isotope measurements. Dissected sample tissues were too heavy for the laser-induced sample transfer that comes with the Zeiss system. The Zeiss system has an automatic z-focus feature that allows easy definition of the number of automatic cutting iterations. For each cutting cycle the focus of the UV-laser beam is adjusted by a pre-defined z-focus delta value. As the Zeiss laser is less powerful than the laser from Leica, more cutting iterations are generally required. Together with the manual collection of dissected specimens, the overall sampling process with the Zeiss system is longer than with the Leica system. However, the dissection process with the Zeiss can be operated automatically, for example overnight, whereas the Leica system requires the full-time presence of the user: first, the adjustment of z-focus of the laser beam has to be done manually; and second, dissected specimens sometimes tilt, stick and do not fall down into the collection vessels as the cutting line of the laser beam is much narrower than the thickness of the sample.

The use of UV-laser microdissection microscopes is not necessarily faster than traditional methods for the dissection of wood tissue. In general, the cutting process of selected tissue lasts 1–2 min, depending on the size of the selected area, the thickness of the cross-section, density of the wood material and the UV-laser settings. If further cutting iterations are required, further time is needed. The average sample throughput per 8-h day may vary between 20 and 120 samples. This includes the on-screen selection of area, the automatic UV-laser-based microscopic dissection and the collection of specimens, as well as unpredictable interferences such as stuck specimens (Leica LMD7000). With some modification of the current sample collection methods the Leica system could also be run automatically overnight, which would increase sample throughput drastically.

At this stage, the key advantage of using the novel technique is on-screen selection of areas and ultrahigh-resolution sampling of plant tissue with unprecedented precision. Furthermore, both systems provide electronic documentation of the dissection processes by photo or video sequences, as well as a report of labelled and dissected elements.

Conclusions

We present a novel method for dissecting plant tissue in stable isotope studies by using UV-laser-based microscopic dissection. Using two commercially available UV-laser microdissection microscopes, we prepared high-resolution intra-annual wood samples from various tree species of different climate zones for C and O isotope analyses with conventional IRMS systems.

The UV-laser-based microscopic dissection enables precise selection of relevant plant cells/tissues on video screen by pen. This procedure allowed us to detect that the δ13C contents of xylem cells and woody ray parenchyma of the deciduous species (Q. robur) have the same year-to-year variability, but also reveal offsets that are opposite in sign depending on whether wholewood or cellulose is analysed. Furthermore, we showed that UV-laser-based microscopic dissection facilitates electronically documented sub-seasonal sampling of tree rings with irregular shapes or narrow ring widths. This is a prerequisite for establishing long and continuous high-resolution isotope chronologies for high-quality climate reconstructions. Single cell rows can be sampled and thus short-term climatic events such as droughts or extensive rainfall events (e.g. caused by ENSO) may be related directly. The three applications demonstrate that the use of this new technique enables the user to sample plant tissue at ultrahigh resolution and unprecedented precision.

This new technique opens new ways for studying the environmental information in tree rings nonlinearly integrated over the vegetation period. The combination of direct growth measurements, which provide exact growth rates of tree rings (e.g. punching method (Forster et al., 2000; Rossi et al., 2006), dendrometers) and precisely selected and dissected specimens may provide a better time-match between high-resolution stable isotope ratios and real-time climate data. Comprehensive eco-physiological process studies on stable isotope signal transfer in the arboreal system can be combined with qualitative and quantitative microscopic investigations on anatomical properties. As wood anatomical properties are climate sensitive (Fonti et al., 2010; Liang et al., 2013; Olano et al., 2013), more and more studies are attempting to combine quantitative wood anatomy and stable isotope analysis (Ponton et al., 2001; Battipaglia et al., 2010; De Micco et al., 2012; Rossi et al., 2013). Such investigations may explicitly profit from UV-laser-based microscopic dissection. For instance, detailed assessment of C and O isotope variability in woody parts such as resin ducts, rays, fibres, vessels or parenchyma cells is possible, as they can now be precisely targeted on thin sections by the novel method.

The use of UV-laser-based microscopic dissection systems offers new possibilities in view of relating plant structure with plant functioning (derived from isotope ratios) in studies on vulnerability, resilience and adaption of plants to past and present global change. Multidisciplinary analyses (cell structure, wood density and wood chemistry analyses) on the same wood samples are now possible with a complete recovery of the sample due to the use of thin sections.

Acknowledgements

We are grateful to Heiko Baschek, David Göhring and Sophie Wendler for support in the laboratory, to Andreas Hendrich for help with the layout of figures, and to Sonia Simard, Wei Liang, Hagen Pieper and Uwe Schollän for fruitful discussions. Furthermore, we thank three anonymous reviewers and the editor for their valuable comments and suggestions. This study was funded by the ISOWOOD-Breeding (BMBF, 0315427B), INDOPAL (HE3089/1-1) and CADY (BMBF, 03G0813H) project.

Ancillary