NMR set-up and flow-imaging methods
NMR imagers that are used for plant work often have been developed for other purposes, such as medical imaging. As a result, many machines consist of horizontal bore superconducting magnets with cylindrical imaging gradients. For plant work this is not an ideal set-up. In such a system, plants have to be placed horizontally instead of vertically, and fitting the shoot or roots of a plant through the narrow cylindrical bore of a gradient set can be stressful and damaging for the plant. In the current study, we used an NMR imager consisting of an electromagnet with an open structure. The imaging gradient set in the centre of the magnet was made up of two flat plates instead of the more conventional cylindrical gradient sets. This configuration allowed us to easily place large plants, up to a size of 2 m, upright in the NMR magnet (Fig. 1). Placement of plants could be undertaken without causing much stress to the subject, other than the stress that is caused by moving and handling it. Potentially stressful actions like mounting an RF coil (Fig. 1b) or, when necessary, the removal of a branch or leaf could be undertaken well in advance. In this study, the plants were left to acclimate to the conditions in the magnet for a minimum of 2 d before commencing measurements, although this may not have been necessary because the plants were growing vigorously and showed uninhibited water uptake within a few hours after they were placed in the magnet.
Measuring phloem flow by means of NMR flow imaging has, from a technical point of view, for a long time remained a very challenging enterprise. Köckenberger et al. (1997) were the first to use NMR flow imaging to measure phloem flow in an intact castor bean seedling grown in the dark. In order to measure flow, they used a ‘difference propagator’ technique that yielded flow profiles that were, as far as the general principle is concerned, comparable to the flow profiles presented in the current study. In order to quantify these (distinctly non-rectangular) flow profiles, the difference propagators were fitted with a rectangular function. This is a valid approach when the following conditions are met: the flow within vessels is laminar; only one vessel is present per pixel, or more vessels are within one pixel but all have the same diameter; and vessels are never only partially within one pixel and partially within another. Under these conditions a flow profile would always have a rectangular shape (broadened by diffusion). However, in practice the recorded flow profiles were not rectangular, but looked like the flow profiles presented in this paper (Figs 2, 5 & 8), showing a large contribution of water moving at lower speeds and a smaller contribution of fast moving water. Fitting a flow profile like this with a rectangular function may yield a correct value for the total volume flow, but the average velocity of the phloem sap would be overestimated. In addition, during data processing, information with regard to the amount of stationary water per pixel is lost.
A more rapid flow-imaging method called FLASH was developed by Rokitta et al. (1999), which was later shown to be capable of measuring phloem and xylem flow in 40-day-old castor bean plants, placed horizontally in a high-field (7 T) superconducting magnet (Peuke et al. 2001). In this approach the spatial resolution was reduced in conjunction with a faster flow-encoding method in order to shorten the measurement time. While the method was quantitative in that it yielded an overall average velocity per pixel, it did not yield information about the velocity profile that gave rise to a particular average velocity, nor did it allow the volume flow to be quantified in absolute units because the flow-conducting area per pixel was not known.
The NMR flow measurement methods presented in the current paper are especially useful in that they provide, for the first time to our knowledge, a means to record quantitative flow profiles (propagators) of flowing phloem water, while at the same time providing a good spatial resolution at acceptable measurement times. The quantitative flow profiles are unique in that they make it possible to calculate the flow-conducting area per pixel, the average flow velocity per pixel, the amount of stationary water per pixel, and the volume flow per pixel, quantitatively and without the necessity to make any assumptions regarding the flow profile (Scheenen et al. 2000b). As a result of the more favourable anatomy of xylem tissue (large vessels, large amounts of flowing water, and relatively high flow velocities), the SNR of a single PFG-SE-TSE xylem flow measurement was already sufficient to construct quantitative flow maps with a good spatial resolution (Figs 3 & 4). To measure phloem transport, a more elaborate strategy had to be employed. It was usually necessary to lower the spatial resolution of PFG-STE-TSE phloem flow measurements, take more averages per individual measurement, and average three or more individual PFG-STE-TSE phloem flow measurements to raise the SNR sufficiently to allow the measurement to be evaluated on a per-pixel basis and calculate the quantitative phloem flow maps shown in Figs 3 and 4. This approach typically required a minimum of 45 min effective measurement time. Alternatively, the information originating from all flow-containing pixels in an image was summed into a one-dimensional flow profile (Fig. 2), making it possible to analyse measurements individually and lowering effective measurement time to 15–30 min. The combination of the NMR imaging set-up and NMR flow-imaging methodology presented in the current paper thus allowed us to routinely measure quantitative flow profiles of both phloem and xylem flow, in a variety of large plants up to a size of 2 m. The ability to measure phloem flow has already found further application in a study of the effects of cold girdling on phloem mass flow in castor bean (Peuke, Windt & Van As 2006).
How fast does phloem sap flow?
We recorded an average phloem flow velocity of 0.25 ± 0.03 mm s−1 in castor bean during the day (Fig. 7b). This value is in good agreement with the phloem flow velocities reported in earlier studies. Hall et al. used 14C carbon labelling to measure phloem flow in 6-week-old castor bean plants, reporting an average velocity of 0.233 mm s−1 (Hall, Baker & Milburn 1971). Grimmer (1999) used 11C carbon labelling and found phloem flow velocities between 0.387 and 0.41 mm s−1. Köckenberger et al. (1997) found an average phloem flow velocity of 0.58 mm s−1 in 6-day-old castor bean seedlings that were grown in the dark, but because of the reasons discussed earlier, this value will most likely be an overestimation. Peuke et al. (2001) used an NMR flow-imaging method (FLASH) to measure phloem flow velocity, and reported a flow velocity of 0.25 mm s−1, averaged over all plants in the experiment.
Remarkably, the flow velocities we measured in poplar, castor bean, tomato and tobacco were all within the same range, irrespective of plant size and species. The highest average linear velocity that was measured was 0.44 mm s−1 in tomato; the lowest recorded value was 0.25 mm s−1 in castor bean (Fig. 7b). The phloem flow velocities in a wide range of plants, as far as these are available from literature, are roughly within the same range as well. In castor bean, phloem flow velocities with values between 0.233 (Hall et al. 1971) and 0.58 mm s−1 were found (Köckenberger et al. 1997). Mortimer (1965) investigated phloem transport in sugar beet petioles using 14C labelling and found phloem flow velocities between 0.139 and 0.417 mm s−1. In soybean, Fisher (1978) measured phloem flow velocities between 0.13 and 0.15 mm s−1. Hartt (1967) reported flow velocities between 0.17 and 0.22 mm s−1 in sugar cane. Taking into account that these flow velocities were measured in different species and using different techniques, the differences are remarkably small. As far as we are aware, only one example – measured in a monocot, in contrast to the studies mentioned before – is an exception. Passioura & Ashford (1974) measured phloem flow velocities of 0.2 up to 1.7 mm s−1 in wheat. However, these values were found in root systems that were pruned to produce exceptionally high flow rates.
It is widely accepted that phloem sap flow is governed by the solute consumption of the terminal and axial sinks, if the sources are strong enough to meet the demand (Thompson 2006). Local volume flow would thus depend on phloem solute content, and sink consumption downstream. The local sap flow velocity would then be determined by the number and the diameters of the phloem conduits that conduct this volume of flow. It would seem fair to assume that in different species, the number and diameter of the phloem conduits in a cross section of stem show considerable variation. The balance between source strength and sink demand may be different as well. This could easily result in large differences in flow velocity. Then why are the average phloem flow velocities that have been measured so far this similar?
One might speculate that the phloem is scaled and regulated to maintain a constant and relatively slow flow of sap. A limitation of sap flow velocity may be necessary, for instance, because too fast a flow would threaten to dislocate parietal proteins or organelles that are present in the sieve tube (Ehlers, Knoblauch & van Bel 2000; van Bel 2003). It is also conceivable that the flow velocity should not be too low, or too variable. If the flow velocity would be too low, for instance, during times of low sink demand, then phloem-borne molecular signals might take too long to arrive at their destinations. When phloem sap flow would be highly variable, then the transit time of molecular signals might become unpredictable.
Diurnal variation in phloem and xylem flow rates
As would be expected, xylem flow in the four plant species exhibited clear diurnal dynamics with regard to flow profile, average volume flow and average linear velocity (Figs 8–10). Interestingly, the decrease in volume flow at night was not only associated with a decrease in average linear velocity, but, for all four plant species, also with a decrease in flow-conducting area (Fig. 10a). The reductions in flow-conducting area were surprisingly large, with values in the order of 50% for castor bean and tobacco, and in the order of 25% for poplar and tomato. It is well known that plant stems can exhibit a diurnal pattern of micro expansion and contraction. However, a contraction of the stem (and perhaps of xylem vessels) would be expected during the day when xylem pressures are most negative, not during the night. Therefore, the reduction in flow-conducting area must have been caused by a reduction in the number of flow-conducting xylem vessels. The slowest moving water will be found in the vessels with the smallest diameters. It is likely that, under conditions of low transpiration, the flowing water in these vessels will be the first to become indistinguishable from stationary water.
A number of early studies have demonstrated that phloem bulk transport is sensitive to changes in plant water status. Hall & Milburn (1973) observed that phloem exudation in castor bean decreased upon the application of water stress, and increased again when the water stress was lifted. Peel & Weatherly (1962) measured diurnal sap exudation rates in rooted willow cuttings. They measured an increase in sap exudation rate in the dark relative to that in the light, showing that not only water stress, but also more subtle differences in plant water status as caused by the differences in day/night transpiration rate, can cause differences in phloem exudation rate. Peuke et al. (2001), in contrast, did not find significant differences between day and night phloem flow velocities that were measured by means of NMR imaging in 35–40-day-old castor bean plants.
In our study, we did not observe significant differences between the day and night-time phloem volume flow rates in any of the plants (Fig. 7c). In tobacco, castor bean and tomato, the day and night flow profiles were virtually identical (Fig. 5), although in tomato the flow-conducting area, average flow velocity and volume flow appeared to be slightly (but not significantly) higher during the day than during the night period (Fig. 7). Poplar was the only plant in which the day- and night-time phloem flow profiles were clearly different (Fig. 5). At night, the phloem flow-conducting area was significantly larger than during the day (Fig. 7a). The increase in flow-conducting area at night was compensated for by a significant decrease in the average flow velocity, so that the resulting volume flow at night remained almost unchanged (Fig. 7c). The increase in flow-conducting area at night could have been caused by an increase in the number of flow-conducting sieve tubes, or by an increase in sieve tube diameter.
Trees exhibit a diurnal pattern of trunk shrinkage and expansion. Up to 50% of this diameter variation was shown to originate from shrinkage and swelling of phloem and bark tissue (Sevanto et al. 2002). In Monterey pine, measurements were taken on trees from which the bark was removed, leaving the phloem tissue uncovered. In this condition, the daily change in thickness of the phloem layer was about 6%. In castor bean, Kallarackal & Milburn (1985) reported that the diameter of stems decreased when phloem turgor was released, by cutting the phloem above or below the site where stem diameter was measured. These reports demonstrate that the phloem can expand and contract. However, in the current study we observed an increase of flow-conducting area in polar, from 2.6 mm2 during the day, to 3.8 mm2 at night. This is an increase of 46%. According to Thompson & Holbrook 2003, the change in cross-sectional sieve tube area as a function of pressure is given by the following formula:
Here, a is the flow-conducting area, p pressure, and ɛ the drained pore elastic modulus. If for sieve tubes an elastic modulus of 30 MPa is assumed (Holtta et al. 2006), then it can be calculated that for this large an increase in flow-conducting area, a change in turgor pressure of about 11 MPa would be required. This pressure difference is much larger than would normally be expected in the phloem. We conclude that the diurnal change in flow-conducting area cannot be explained by shrinkage and expansion of sieve tubes, but is probably caused by a change in the number of flow-conducting conduits.
It is known that diurnal sucrose export rates can vary considerably, within a diurnal cycle as well as between plants. Grodzinski, Jiao & Leonardos (1998) used 14CO2 labelling to determine the daytime carbon export rate in 21 plant species. At ambient CO2, they found linear relationships between photosynthesis, sugar synthesis and concurrent export. At high CO2, when more sugars were assimilated, the relationships between photosynthesis and export rate and between sugar synthesis and export rate were weaker, probably because the phloem export capacity became limiting and sugars and starch were accumulated in the leaf.
Grimmer & Komor (1999) measured the phloem transport rate in castor bean plants grown under elevated and normal CO2 conditions by pulse labelling with 11CO2. During the daytime period, they did not find a higher velocity or sucrose concentration in the sieve tube sap from plants under elevated CO2, despite the fact that these plants were assimilating more carbon. The carbon balance of the source leaves indicated that the carbon export rate during the day was the same for both CO2 conditions. However, large differences showed up at night. The carbon export rate of plants grown at normal CO2 declined to approximately half the daytime rate, while the carbon export in plants grown under elevated CO2 remained high. Furthermore, the plants grown at elevated CO2 accumulated starch in the leaves, in contrast to the plants grown at normal CO2. It was concluded, based on these and other findings, that the carbon export rate is at its upper limit in a series of ambient and experimental conditions (Komor 2000). Under ambient CO2 conditions, plants probably operate at or near their maximum carbon export rate during the day period, and at a lower carbon export rate at night, when the carbon pools in the leaf are slowly drained.
Because sucrose is the main osmoticum driving the phloem transport stream (van Bel 2003), the phloem bulk flow might be expected to slow down concurrently with a decline in carbon export at night. This, however, was not observed in the current study. The average phloem volume flows in poplar, castor bean, tomato and tobacco showed small fluctuations throughout the diurnal cycle (Fig. 6), but averaged over a whole day or night period the differences were insignificant (Fig. 7). In the current study, the carbon export rate and sucrose content of the phloem sap were not measured, so we cannot conclude directly that the phloem flow was kept constant independently from declining carbon export rates. However, our results correspond nicely with the observations published by Peuke et al. (2001), who found that sucrose concentrations in the phloem sap dropped by 12% at night, while the day/night phloem flow velocities remained unchanged.
These results support the idea that phloem flow velocities in plants are conservative in nature, and that phloem sap flow may be regulated to remain constant, regardless of changes in apoplastic water potential, source strength, or sink solute consumption. One might speculate that phloem volume flow is regulated by modulating phloem solute content. This might be achieved at the sources or the terminal sinks, but also along the length of transport phloem, where a rigorously regulated process of release and retrieval of photosynthates has been shown to take place (Aloni, Wyse & Griffith 1986; Minchin & Thorpe 1987; van Bel 2003).
Water flowing upward in the xylem is replenishing water that is used by three main processes: transpiration, growth and phloem mass flow. The latter process is known as recirculation (Pate et al. 1985) or Münch’s counterflow (Tanner & Beevers 1990, 2001) It is commonly estimated that the contribution of Münch’s counterflow to the total xylem volume flow in transpiring plants is very low or even negligible (e.g. Jeschke et al. 1996). Our results show that during the day the phloem to xylem ratios indeed were low (although not completely negligible), with a maximum value of 0.10 (10%) in tobacco. However, during the night, the phloem to xylem ratio increased to much higher values. In tobacco and castor bean values of 0.54 and 0.37 were measured, respectively, implying that in tobacco 54% and in castor bean 37% of the xylem bulk flow at night is actually generated and recirculated by the phloem.
The recycling of xylem water by way of the phloem was measured for the first time by Köckenberger et al. (1997), in a 6-day-old castor bean seedling grown in the dark. In the stem of the seedling, just below the cotyledons, they measured a xylem volume flow of 38 µL h−1 (5 µL h−1 of which was estimated to be used for growth) and a phloem volume flow of 17 µL h−1, giving a phloem to xylem ratio of 0.45 – a value close to the one we measured in a full-grown castor bean plant during the dark period (Fig. 11). This is surprising because a seedling, in contrast to a fully developed plant, is not expected to transpire much. In the seedling, the phloem to xylem ratio would therefore be expected to be higher than in the full-grown plant. Jeschke et al. (1996) modelled water flows in full-grown 44–53-day-old castor bean plants. They estimated that in a 9 d period, only 1% of water was transported downward in the phloem compared with the upward directed transpiration stream in the xylem. Even though no distinction was made between day- or night-time transport, this estimate appears to be too low.
Tanner & Beevers (2001) investigated the question whether transpiration is required to supply the plant with minerals. They did so by growing sunflower plants under two conditions, one where mineral nutrition was only supplied during the 12 h day period, and one where it was only supplied during the 12 h night period while the plants grew under conditions of near 100% relative humidity. They did not find any difference in the growth rates of both groups of plants, indicating that plants were able to take up and distribute minerals in the absence of transpiration. Making a number of assumptions, they estimated that the contribution of Münch’s counterflow would have been 400 mL out of a total volume of 1985 mL, giving a phloem to xylem ratio of 0.20. This estimate compares well with the phloem to xylem ratios that we present in the current paper.
We conclude that throughout the day, and especially at night, a significant percentage of xylem water is not transpired but recirculated by means of the phloem. Münch’s counterflow may thus play a significant role in maintaining xylem circulation at night.