JXB Advance Access originally published online on April 4, 2005
Journal of Experimental Botany 2005 56(415):1419-1425; doi:10.1093/jxb/eri143
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RESEARCH PAPER |
Input–output analysis of in vivo photoassimilate translocation using Positron-Emitting Tracer Imaging System (PETIS) data
1Takasaki Radiation Chemistry Research Establishment, Japan Atomic Energy Research Institute, Gunma 370-1207, Japan
2Department of Radioisotopes, Japan Atomic Energy Research Institute, Gunma 370-1207, Japan
3Central Research Laboratory, Hamamatsu Photonics Co., Shizuoka 434-0041, Japan
* Present address and to whom correspondence should be sent: Institute of Agricultural Chemistry, Carl-Sprengel-Weg 1, D-37075 Göttingen, Germany. Fax: +49 551 395570. E-mail: Akeutge{at}gwdg.de. For correspondence on PETIS, please contact Dr S Matsuhashi; E-mail: Shinpei{at}taka.jaeri.go.jp
Received 27 October 2004; Accepted 23 February 2005
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionReferences |
|---|
The Positron-Emitting Tracer Imaging System (PETIS) is introduced for monitoring the distribution of 11C-labelled photoassimilates in Sorghum. The obtained two-dimensional image data were quantitatively analysed using a transfer function analysis approach. While one half of a Sorghum root in a split root system was treated with either 0, 100, or 500 mM NaCl dissolved in the nutrient solution, tracer images of the root halves and the lower stem section were recorded using PETIS. From the observed tracer levels, parameters were estimated, from which the mean speed of tracer transport and the proportion of tracer moved between specified image positions were deduced. Transport speed varied between 0.7 and 1.8 cm min–1 with the difference depending on which part of the stem was involved. When data were collected in the lowest 0.5–1 cm of the stem, which included the point where the roots emerge, transport speed was less. Rapid changes in NaCl concentration, from 0 to 100 mM, resulted in short-term increases of assimilate import into the treated root. This response represented a transient osmotic effect, that was compensated for in the medium-term by osmotic adaptation. Higher concentrations of NaCl (500 mM) resulted in distinctly less photoassimilate transport into the treated root half. The present results agree with earlier observations, showing that transport of 11C-labelled photoassimilates measured with the PETIS detector system can be quantified using the method of input–output analysis. It is worth noting that with the PETIS detector system, areas of interest do not need to be defined until after data collection. This means that unexpected behaviour of a plant organ will be seen, which is not necessarily the case with conventional detector systems looking at predefined areas of interest.
Key words: 11C carbon, short-lived radioisotope, transfer function analysis
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
In vivo studies of phloem transport in plants have frequently used the short-lived positron-emitting radioisotope 11C with its half-life of 20.4 min (Minchin and Thorpe, 2003
; Fujikake et al., 2003; Keutgen et al., 1995; Roeb and Britz, 1991). Detection of the radioisotope distribution in plants has mostly been performed using NaI scintillation detectors, either arranged in pairs for coincidence measurement or very well shielded with lead or tungsten to ensure that they are sensitive only to radiation from a well-defined part of the plant. These detector techniques are useful to follow the 11C-movement between two positions of a plant (Minchin, 1978; Thorpe et al., 1983) or to measure the relative distribution of 11C-labelled carbohydrates within the entire plant (Williams et al., 1991; Thorpe and Minchin, 1991). These have been referred to as ‘slot’ and ‘sink’ modes of detection (Minchin and Thorpe, 2003). In the ‘slot’ mode, tracer movement is observed within a short segment of leaf or stem through a slot in the radiation shielding. Because the tracer is measured only within a tissue segment, quantitative data interpretation is difficult (Minchin and Thorpe, 1987). For quantitative analyses the ‘sink’ mode of detection is preferred, with tracer measured within a terminal sink region with uniform sensitivity. Because tracer export from a terminal sink region is close to zero, quantitative calculations using the method of input–output analysis are possible on time-varying systems (Minchin and Thorpe, 2003). Its application is not limited to NaI detectors, and may also be used with other detector types.
The Hamamatsu Photonics Co. has developed, in co-operation with the Japanese Atomic Energy Research Institute (JAERI) in Takasaki, Japan, an alternative two-dimensional imaging system that is able to detect positron-emitting tracers in living organisms with an improved spatial and temporal resolution. This positron-emitting tracer imaging system (PETIS) enables to study biological transport processes in plants in real time in an area of approximately 50x150 mm2 with a sampling interval of as low as 5 s and spatial resolution of 2 mm. The limit to the spatial resolution to detect where an 11C-decay occurs is the 0.8 mm average path length of the emitted positron in wet tissue. However, if an 11C atom decays close to the plant surface, the emitted positron can travel up to 4 m in air before it annihilates. Because phloem tissue is often close to stem or leaf surfaces, detection sensitivity can be increased by using ‘positron shields’ close to the plant surfaces in order to ensure that annihilation of the positrons occurs as close as possible to its site of production (Minchin and Thorpe, 2003). However, this reduces the spatial resolution, as it forces a positron that had escaped the plant tissue to annihilate in the shield, leading to the observation of tracer within a larger detection volume, which results in an increased blurring of the image. The properties of positrons and plants as well as shielding to increase detection sensitivity limits the spatial resolution of PETIS to about 2 mm.
The advantage of PETIS over conventional detection is its ability to detect radiation emitted from entire (small) plants or larger plant parts. Within the view of the detector, regions of interest can later be selected freely, and transport between these regions can be statistically analysed using input–output analysis. The most important factor limiting the size of selected areas is the amount of radioactivity coming from these areas, which must be significantly above the statistical noise level. It is necessary to ensure that the detector is equally sensitive to tracer anywhere within the plant, which is best achieved by introducing sensitivity corrections.
An aim of the present study was to demonstrate, for the first time, that transfer function analysis is possible for tracer data collected with PETIS. For example, it was decided to look at the effect of NaCl salinity in the root system of Sorghum on assimilate partitioning using the split root system of Williams et al. (1991). The advantage of a split root system is the possibility of comparing the partitioning of labelled carbohydrates into two morphologically identical sinks, supplied along an identical length of vascular tissue from a common source leaf, when one root-half is exposed to a stress and the second one is untreated and serves as an internal control (Williams et al., 1991). Williams et al. investigated the short- and medium-term influence of osmotica in the root medium on assimilate partitioning in barley. As NaCl acts as osmoticum, similar responses were expected in these experiments, allowing an evaluation of the advantages and disadvantages of PETIS.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
Plant material and growth conditions
Seeds of Sorghum bicolor (L.) Moench cv. Fela bodedio were surface-sterilized by soaking in 70% ethanol solution (v/v) for 1 min, then in 0.5% sodium hypochlorite solution (w/v) for 3 min. After a thorough rinsing in sterile deionized water, seeds were germinated on filter paper in the dark at 25 °C. Five days later, the seedlings were transferred to 5.0 l plastic containers (10 seedlings per container) filled with vermiculite and supplied with modified Hoagland solution (Sruamsiri and Lenz, 1985
). After 3 weeks preculture, they were hydroponically grown in two 8 ml glass vessels and the root of each plant was split equally between the two vessels. A modified quarter-strength Hoagland nutrient solution was supplied and exchanged every day during the last 2 or 3 weeks. When hydroponically grown, a slightly different type of root developed, which was used for the experiments. Older roots were subsequently removed. Growth conditions in the controlled environment chamber were a 16 h photoperiod, a 26/24 °C day/night temperature regime, and a relative humidity of about 70%. Light intensity was about 350 µmol m–2 s–1 photosynthetically active radiation at plant height. Sorghum plants were used for the 11C-experiments when about 6 weeks old and about 20–25 cm in height. They were arranged in the experimental set-up 2 d before 11CO2 application to allow time for adaptation (Figs 1, 2).
|
|
Synthesis of 11CO2 and PETIS measurements
The positron-emitting 11C-radioisotope was produced by bombarding a nitrogen gas target with 1 µA of 10 MeV protons from the TIARA AVF cyclotron (Arakawa et al., 1995
). The 11C-atoms reacted with traces of oxygen in the nitrogen gas to form 11CO2, which was collected and enriched, transferred to the controlled environment chamber, and supplied to the penultimate Sorghum leaf using a leaf cuvette. The 11CO2 was assimilated by photosynthesis for 5 min (pulse application) and incorporated into leaf carbohydrates, which were distributed within the plant. The 11CO2-enriched air in the cuvette was replaced after 5 min by non-labelled air. Each 11C-translocation study lasted for 90 min. The distribution of tracer was followed with PETIS. This system consists of two-dimensional block detectors (50x150 mm2), composed of Bi4Ge3O12 scintillator arrays coupled with a position sensitive photomultiplier tube (PS-PMT, Hamamatsu R3941–2, Hamamatsu Photonics Co., Hamakita, Shizuoka, Japan), using coincidence detection of the two annihilation 
-rays (Kume et al., 1997a, b).
A single Sorghum plant was placed midway between a pair of PETIS detectors, so that the two halves of the root within the two glass vessels and the lower 9 cm of the shoot, which was placed between two acrylic boards, were within the view of the detector system. 11C-radioactivity remaining in the application area of the leaf was not detected by PETIS, although it usually accounts for more than 50% of the applied tracer activity. The 2 mm thick acrylic boards served as ‘positron shields’ to improve detection sensitivity. These were placed about 2 mm from the stem surface to allow gas exchange. The glass of the root vessels and the nutrient solution had a similar trapping effect for positrons escaping the root tissue, although the -absorption would be slightly different in the root area (glass vessel+nutrient solution) compared with the stem region (acrylic board), but this can be taken care of by appropriate sensitivity corrections. For the young Sorghum plants used during the experiments, the term ‘stem’ must be used with reservation, because a real stem did not develop. The older leaves surrounded the newer ones and all originated at the base of the shoot, known as the ‘discrimination centre’ (Mori, 1998). Nevertheless, for convenience, the term ‘stem’ is used in the present article, although it might correctly be referred to as basal leaf sheath.
During the 11C-experiments, a time-series of two-dimensional images of 11C-labelled carbohydrates distribution in Sorghum was collected for 90 min, with a sampling interval of 30 s. The left root section was always exposed to 8 ml modified quarter-strength Hoagland nutrient solution, and the right section to either this solution (control) or NaCl at either 100 or 500 mM prepared by dissolving NaCl in 8 ml modified quarter-strength Hoagland nutrient solution.
During the first control experiment, the effect of exchanging nutrient solution on assimilate partitioning between the root halves was investigated. 11CO2 was supplied to a single Sorghum leaf three times for 5 min at 2 h intervals. First, the nutrient solution within both glass vessels was exchanged 4 h before the first tracer application. Immediately after the first measurement, about 30 min before the second tracer application, the nutrient solution in both glass vessels was once again replaced by 8 ml modified quarter-strength Hoagland nutrient solution. Similarly, about 30 min before the third tracer application, the nutrient solution surrounding only the right root half was replaced. In order to minimize plant handling, solution changes were performed using a needle connected to a 10 ml syringe via a thin plastic tube. The experimental procedure took advantage of 11C's short half-life of 20.4 min, allowing multiple labelling of the same plant. Hence, a single plant may serve as its own control.
For the second experiment, 11CO2 was supplied to Sorghum three times under the same time schedule. The first application served as a control to quantify the distribution of 11C-labelled carbohydrates into the root halves without NaCl stress. Thirty minutes before the second and third applications, the nutrient solution within the vessel containing the right root half was replaced by 8 ml modified quarter-strength Hoagland nutrient solution with 100 mM NaCl.
During the third and fourth experiments, the nutrient solution around the right root half was exchanged as described for the second experiment, but this time 500 mM NaCl was applied for the third application. This elevated concentration was chosen to investigate the plant's short-term response to a toxic salt level.
Quantification of PETIS data
Pulse experiments with 11C as radioactive tracer have frequently been analysed by a decay correction of the observed tracer profiles. In the present experiment with multiple tracer loading, this procedure was no longer possible, because tracer observed at a given time could have taken a range of times since loading to reach the observation site. Minchin (1978, 1979) and Minchin and Troughton (1980) proposed a method of analysis, based upon finding a transfer function, and which incorporates radioactive decay as a part of the tracer dynamics rather than requiring the data to be decay corrected. Changes in the tracer profiles along the transport pathway are described quantitatively by input–output equations, fully describing the system's dynamics. The data measured with Sorghum plants were well described by the following input–output relationship (Minchin and Troughton, 1980):
 | (1) |
and 
(Minchin and Thorpe, 1989) by
 | (2) |
 | (3) |
is the decay constant for the tracer. Parameters 
and 
are estimated from the observed tracer levels, and the decay corrected parameter values a1 and bj are then calculated from equations (4) and (5):
 | (4) |
 | (5) |
; Minchin and Troughton, 1980) by the equation:
 | (6) |
) by the equation:
 | (7) |
). Finally, the model with the smallest model error variance was selected. Various model structures were evaluated and, once the ‘best’ model was found, it was accepted or rejected on the basis of known information: for example, a comparison of the relative decay-corrected radioactivities in the ‘stem’, ‘discrimination centre’ and root sections at the end of a single pulse experiment, which can be regarded as an approximation of the translocated tracer activity.
A problem in data analysis arises from the different measurement sensitivities for the various plant parts within the field of view brought about by the small differences in distance from the radiation detectors, as well as the different surrounding mediums for the roots and ‘stem’. However, by treating the two root halves and the ‘stem’ as lumped systems, the ratio of partitioning between the root halves could be calculated. Comparison of tracer partitioning within each of the four experiments is possible, as the experimental set-up was not changed during the three treatments leaving the geometry unchanged. Moreover, the calculation of average transit times is comparatively insensitive to unequal detector sensitivities (Minchin, 1979). Therefore, it was possible to calculate the mean transport speed of 11C-labelled assimilates in the ‘stem’ up to the ‘discrimination centre’ (Speed 1) and up to the root halves, including the ‘discrimination centre’ (Speed 2 and Speed 3). Speed 2 and 3 represent the mean transport speeds of 11C-labelled assimilates transported into the right and left root halves, respectively. The mean transport speed was calculated by dividing the length of the transport pathway (8.5–9.5 cm) by the average transit time t.
The quantitative analyses of the tracer profiles is based on the changes in the shape between the input and output data, described by the input–output equations. The input data were given by the entire radioactivity seen in the areas 1+2+3+4 of the PETIS detectors (Fig. 3). For the left and right root-halves, output was either the radioactivity in area 3 or area 4 (output 2/3 in Fig. 3). Transport speed to the roots (Speed 2 and Speed 3) was measured accordingly. Transport speed within the ‘stem’ up to the ‘discrimination centre’ (Speed 1) was calculated by taking the radioactivity seen in the areas 1+2+3+4 as input and those of areas 2+3+4 (‘discrimination centre’ and roots) as output (output 1 in Fig. 3).
|
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
With PETIS, the two-dimensional time-course of tracer movement through the ‘stem’ and into the roots can be visualized in vivo. Figure 4 shows the data for one such measurement, made on a non-stressed Sorghum with a split root. The root mass on the right side was larger than that on the left. About 12 min after 11CO2 pulse labelling, tracer first appeared in the upper ‘stem’ viewed by PETIS (area 1 in Fig. 3). The ‘discrimination centre’ (area 2 in Fig. 3), which was characterized by the accumulation of tracer during the measurement, received tracer at about 21 min and by 24 min was seen to be accumulating tracer above the level of the nearby vascular tissue. Tracer first appeared in the roots soon after it arrived at the ‘discrimination centre’ by 26 min.
|
The speed of tracer movement through the ‘stem’ above the ‘discrimination centre’ was 1.8±0.6 cm min–1 (means ±standard deviation, n=12), which is similar to that measured in maize leaves (Minchin, 1979
), another C4 graminaceous species. When the ‘discrimination centre’ was included, the average transit speed was more than halved (0.7±0.1 cm min–1, n=23), which is consistent with the observation that the ‘discrimination centre’ accumulates tracer and releases it back into the transport stream. It is a localized short-term storage zone. The ‘discrimination centre’ must be releasing the tracer within the time scale of these measurements or else the release would not be seen and the ‘discrimination centre’ would not alter the transport speed, but just the fraction of the tracer that moved through this region. It is clear from the PETIS images that not all the tracer that entered into the field of view was transported into the roots, a fraction remained within the plant ‘stem’ and also within the transport pathways associated with the roots. This is a pictorial representation of phloem loss along the length of the transport phloem, which is quantified by the suggested method.
The PETIS images are readily quantified by defining areas of interest and extracting the time-course of tracer levels within these areas. Detailed analysis of these tracer profiles gave the partitioning fraction for each root half, which is the fraction of the available tracer that eventually enters each root half. These fractions are shown in Fig. 5, experiment 1, for three tracer pulses applied to the same plant. During this control experiment, exchanging the nutrient solution did not influence assimilate partitioning between the root halves. Only import into the entire root was slightly reduced during the course of the day, which is consistent with earlier observations (Williams et al., 1991).
|
When one root half of a split root was transferred into 100 mM NaCl, the PETIS images (not shown) did not indicate any clearly visible difference to the control measurement made on the same root with the previous tracer pulse. Quantification of tracer transport into the roots, however, clearly showed an increase in partitioning to the treated root half (Fig. 5, experiments 2, 3, 4) soon after NaCl was applied. However, by the third tracer pulse (Fig. 5, experiment 2), 2 h later, root partitioning had return to its pretreatment value. When 500 mM NaCl as applied by the third tracer pulse (Fig. 5, experiments 3, 4) partitioning to the treated root declined to well below that of the first control tracer pulse. This strong effect was clearly visible in the PETIS images (data not shown).
The response seen when using 100 mM NaCl is in full agreement with osmotica results reported by Williams et al. (1991). The immediate increase in partitioning to the treated root half represents a direct consequence of lowering the water potential of the treated root, which caused an increase in phloem flow into this root. By the time of the next tracer pulse, 2 h later, the import into the Sorghum root treated with 100 mM NaCl (Fig. 5, experiment 2) was no longer elevated. This reflects a medium-term increase of the water potential of the treated root half due to osmoregulation. The response to the high salinity level (500 mM NaCl), on the contrary, is interpreted as a toxic response (Fig. 5, experiments 3, 4). A similar experiment was reported in Wieneke and Fritz (1985) and Fritz et al. (1987), who supplied a highly toxic NaCl concentration to soybean plants, which resulted in a drastically reduced photosynthesis rate and a permanent reduction in transport of 11C-labelled assimilates to the NaCl-exposed roots. In Sorghum, a reduction of net photosynthesis rate by c. 20% 15–30 min after the application of 500 mM NaCl, which was at least partly due to stomata closure (Netondo et al., 2004), was also observed (AJ Keutgen, unpublished results) in addition to the decline of tracer import. This indicates that changes in assimilate partitioning at toxic NaCl levels may not only be due to a modification of sink but also of source attributes.
It has been shown that PETIS data provide a pictorial view of tracer movement within a plant, and that quantitative analysis involving areas of interest is possible to extract details about the relative amounts and speeds involved. The spatial resolution of PETIS is well demonstrated and the advantages of not having to define areas of interest until after data collection. However, the size of PETIS (50x150 mm2) still limits the experiments that are possible. Using PETIS in conjunction with conventional detector methods provides a partial solution. To obtain quantitative data and to correlate the PETIS data with the radioactivity really fixed by the application leaf, future experiments should involve at least two additional NaI detectors: one monitoring the radioactivity of the entire plant, the other monitoring the plant without the application leaf. PETIS detectors are ideal to measure the temporal and spatial distribution of tracer within a (terminal) sink and/or an area of interest between source and sink with a uniform emission of radioactive tracer signals.
|
Acknowledgements |
|---|
The authors would like to express their sincere thank to Dr Peter Minchin for his help on transfer function analysis and his significant input in improving the manuscript. This article is dedicated to him and his colleagues, who introduced input–output analysis for the quantification of 11C-translocation in living plants. The support of Dr Anna and Dr Norbert Keutgen by the Science and Technology Agency (Japan) and the Alexander von Humboldt foundation (Germany) is gratefully acknowledged.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
Arakawa K, Nakamura Y, Yokota W, Nara T, Agematsu T, Okumura S, Ishibori I, Fukuda M. 1995. Status report on JAERI-AVF cyclotron. In: Proceedings of the 14th international conference on cyclotrons and their application, Cape Town, South Africa, 57–60.
Fritz R, Wieneke J, Führ F. 1987. Transport kinetics of 11C-labelled photosynthesis products in soybean. II. Influence of salt stress. Journal of Plant Nutrition 10, 187–205.
Fujikake H, Yamazaki A, Ohtake N, et al. 2003. Quick and reversible inhibition of soybean root nodule growth by nitrate involves a decrease in sucrose supply to nodules. Journal of Experimental Botany 54, 1379–1388.
Keutgen N, Roeb GW, Minchin PEH, Führ F. 1995. Use of transfer function and compartmental analysis to quantify 11C-labelled photoassimilate export from wheat leaves. Journal of Experimental Botany 46, 489–496.
Kume T, Matsuhashi S, Shimazu M, et al. 1997a. Uptake and transport of positron-emitting tracer (18F) in plants. Applied Radiation and Isotpes 48, 1035–1043.[CrossRef]
Kume T, Matsuhashi S, Shimazu M, Ito H, Uchida H, Tsuji A, Shigeta N, Matsuoka H, Osa A, Sekine T. 1997b. Uptake and transport of positron-emitting tracer in irradiated plants. In: Ando T, Fujita K, Mae T, Matsumoto H, Mori S, Sekiya J, eds. Plant nutrition: for sustainable food production and environment. Dordrecht: Kluwer Academic Publishers, 169–170.
Minchin PEH. 1978. Analysis of tracer profiles with applications to phloem transport. Journal of Experimental Botany 29, 1441–1450.
Minchin PEH. 1979. The relationship between spatial and temporal tracer profiles in transport studies. Journal of Experimental Botany 30, 1171–1178.
Minchin PEH, Thorpe MR. 1987. Measurement of unloading and reloading of photoassimilate within the stem of bean. Journal of Experimental Botany 38, 211–220.
Minchin PEH, Thorpe MR. 1989. Carbon partitioning to whole versus surgically modified ovules of pea: An application of the in vivo measurement of carbon flows over many hours using the short-lived isotope carbon-11. Journal of Experimental Botany 40, 781–787.
Minchin PEH, Thorpe MR. 2003. Using the shored-lived isotope 11C in mechanistic studies of photosynthate transport. Functional Plant Biology 30, 831–841.[CrossRef]
Minchin PEH, Troughton JH. 1980. Quantitative interpretation of phloem translocation data. Annual Review of Plant Physiology 31, 191–215.[Web of Science]
Mori S. 1998. Iron transport in graminaceous plants. In: Sigel A, Sigel H, eds. Iron transport and storage in microorganisms, plants and animals, Vol. 35. Metal ions in biological systems. New York: Marcel Decker Inc., 215–237.
Netondo GW, Onyango JC, Beck E. 2004. Sorghum and salinity. II. Gas exchange and chlorophyll fluorescence of Sorghum under salter stress. Crop Science 44, 806–811.
Roeb G, Britz SJ. 1991. Short-term fluctuations in the transport of assimilates to the ear of wheat measured with steady-state 11C-CO2-labelling on the flag leaf. Journal of Experimental Botany 41, 469–475.
Sruamsiri P, Lenz F. 1985. Photosynthese und stomatäres Verhalten bei Erdbeeren. I. Einfluß von Licht. Gartenbauwissenschaft 50, 78–83.
Thorpe MR, Lang A, Minchin PEH. 1983. Short-term interactions between flows of photosynthate. Journal of Experimental Botany 34, 10–19.
Thorpe MR, Minchin PEH. 1991. Continuous monitoring of fluxes of photoassimilate in leaves and whole plants. Journal of Experimental Botany 42, 461–468.
Wieneke J, Fritz R. 1985. Influence of salinity on assimilate partitioning in soybeans. Acta Universitatis Agriculturae Brno, A Facultas Agronomica 33, 653–657.
Williams JHH, Minchin PEH, Farrar JF. 1991. Carbon partitioning in split root systems of barley: The effect of osmotica. Journal of Experimental Botany 42, 453–460.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||