Effect of bark stripping on the electrical impedance of Quercus suber leaves

is study examined the eect of bark stripping on the electrical impedance parameters of cork oak young leaves between 40Hz and 100 kHz. � is was a new application of the electrical impedances spectroscopy (EIS) in plant science. Various stripping coecients (CD) were applied on the trees. Bark stripping is expected to aect water metabolism of leaves and therefore changes in the EIS parameters are expected as well. Single-DCE (ZARC) model was used as equivalent circuit for leaves. Several electrophysiological parameters of this model were compared with moisture content (MC) of the leaves. Intracellular resistance (R i ), extracellular resistance (R e ) and relaxation time (�) of the leaves increased during 14 days aer stripping while the distributed coecient (�) and MC decreased. Signicant correlation between EIS parameters, MC and trees treatments were found.


INTRODUCTION
It is predicted that the cork extraction is extremely traumatic for cork oak trees; it exposes a considerably part of living tissues and causes an extensive wound (Natividade 1950). e e ect of bark stripping is very complex; the large amount of water lost from the stripped surface is most likely to in uence the water balance and, directly or indirectly, limits the physiological processes (Correia et al. 1992). Several methods are used to assess the e ect of stripping intensity on Quercus suber trees e.g. chlorophyll uorescence emission and the stomatal conductance of leaves (Correia 1992, Werner andCorreia 1996), tree diameter growth and its seasonal pattern either in very young cork oak trees (Fialho et al. 2001) or in mature trees under bark production (Costa et al. 2004). Despite numerous studies on Quercus suber trees, a er stripping, there has been made no (or little) attempt to relate the e ect of bark stripping on the dehydratation stress of the leaves.
EIS has been used to study stress reaction in plants (Repo et al. 1994Väinölä and Repo 2000;Laarabi et al. 2005aLaarabi et al. , 2005b. e method provides information about the physicochemical properties of cellular structure. For EIS studies the sample is set in an alternating electric eld with small amplitude of current passing through the sample. When changes in the amplitude and phase of the alternating current are transformed into the complex plane, they induce an impedance spectrum, with frequency as an intrinsic variable. In biological samples, the proportion of current passing through the extracellular 196 spaces (intercellular spaces between the cells) and intracellular spaces in tissues depends on the alternating current (AC) frequency. At low frequency, AC passes through the extracellular space. e cell membranes become more conductive as current frequency increases (Repo et al. 2004) and accordingly, the amount of intracellular current increases. erefore, information about di erent tissues features, i.e., intra and extracellular uids may be revealed by EIS (Väinölä and Repo 2000). Several stressors a ect the EIS parameters of plants organs. e τ and the R i correlate with cold hardening in Scots pine, R e and τ decreases with increases in cellular injuries caused by frost (Repo et al. 1994), the R e correlates with frost-expose in rhododendron leaves (Väinölä and Repo 2000). e stripping, with all its implications in physiological processes (Pinto et al. 2006) may lead to considerable water losses (Correia et al.1992).
Because EIS is sensitive to changes in the physicochemical properties of cells we hypothesized that cellular deterioration in cork oak leaves caused by water balance disturbance a er stripping will alter some EIS parameters as has been observed for other stressors (e.g., Repo et al. 1994Repo et al. , 2000.

Plant material
e study was carried out on young leaves of mature Quercus suber L. trees at site A VII-5 (Canton A, Perimeter VII and Plot 5) of western Mamora forest (6°45'O, 34°2'N, 30m of altitude), Rabat, Morocco. Mamora soil is made of a clay layer on which we have red or pink colour sand layer. e average annual precipitations is 500 mm and the average temperature in summer (June, July, August and September) oscillate around 23°C. Six mature and productive cork oak trees representative of the dominant and typical tree in that eld were selected and numbered T 1 to T 6 . 2 non stripped trees were served as controls (T 1 and T 2 ); two trees (T 3 and T 4 ) stripped for the rst time with CD respectively 1.1 and 1.22, the two other trees (T 5 and T 6 ) bark-stripped one time with CD respectively 1.64 and 1.74. All these trees were regenerated by reject of stump a er ablation of the trunk. Bark stripping characteristics of these trees are outlined in table 1. Table 1. Stripping Characteristics of the trees under study. ND: number of bark-stripping, it is the number of time that the tree was stripped, CD: Bark stripping coe cient calculated as the ratio of the maximum length by over the trunk perimeter at breast height (Natividade 1950

Impedance measurements
EIS of leaves was performed in summer 2009 (before and a er bark-stripping period). ere were ve sampling dates, i.e., on July 7, (few hours before stripping), July 14, (seven days a er stripping), July 22, (fourteen days a er stripping), 7 august (thirty days a er stripping), August 27, ( y days a er stripping). From each tree and at each sampling date, one short shoot was sampled each time from branches located in the central crows of the tree. e leaves were placed in the plastic bags immediately a er sampling and transported to the laboratory. e leaves were oriented vertically when running the impedance measurements. Impedance spectra were measured with an Ag/AgClcell connected to a Hewlett-Packard 3330 LCZ meter. e Ag/AgCl electrodes were kept in contact with the samples using a conductive paste (3M red Dot Foam Monitoring Electrode 2237-3, of the type commonly used for electrocardiograms) to maintain minimum electrode tissue interface polarization. Further, the device was calibrated by using OPEN/SHORT circuit correction to eliminate the e ect of electrodes-paste interface. e real (Z r ) and imaginary (Z i ) values of impedance were then measured within a frequency range of 40 Hz to 100 kHz. e input voltage of the signal was 30 mV (rms). e section of the conductive part of electrodes was 0.78 cm 2 corresponding to a disk of 1cm of as diameter. From each short shoot, 3 young leaves were selected and numbered L 1 to L 3 . L 1, L 2 and L 3 were respectively the rst, the second and the third leaf formed on the short shoot. e three leaves were tested and a leaf thickness was measured with a digital Mitutoyo Calipers 0.01 mm. All the tested leaves had an area of more than 0.78 cm 2 and with naked eye; they presented no sign of aging.

Determination of MC of leaves
A er the impedance measurements, the specimen were weighed and dried at 100°C ±3 for 24h. e scale of 1mg accuracy was used in weighing. e MC of leaves was calculated as: ( 1) Where M H is weight of leaves and M 0 is the oven dry weight a er drying at 100°C for 24h.

Modelling of leaves impedance
Impedance analysis in plants is performed using mainly two types of equivalent circuits i.e., lumped (Harker and Maindonald 1994) and distributed models (Repo et al. 1994, Stout 1988, Mancuso 1999, Mizukami 2007. In this study, the mathematical model ZARC illustrated by the circuit diagram in gure 1 was tted to the data. e ZARC comprises includes a distributed circuit element (DCE) in series with a resistor (R ∞ ). e DCE element comprises includes a Constant Phase Element (CPE) in parallel with a resistor (R). e CPE impedance is de ned as: e real and imaginary parts of impedance obtained from equation (3) are expressed respectively as (see appendix 2 for more explanations): Parameter Ψ is de ning if the center of the arc is below x-axis or not. e angle between x axis and the line passing by the center and the origin of the axes was as and α and Ψ is de ned as: (6) e total complex impedance (Z) of the model of gure 1 is as shown (for derivation of this equation see Appendix 1): By increasing the Ψ value, the impedance spectrum approaches a symmetric arc (attained when Ψ =1). R e was calculated as: R ∞ represents the resistance at high frequency e R i was calculated as: And the relaxation time as: (10) e speci c resistances: intracellular speci c resistance r i (unit Ωm) and extracellular speci c resistance r e (unit Ωm) were obtained by normalization of the measured resistance (R i and R e in respect of the cross-sectional area (A s ) and the thickness of the leaf (e)).
In the text, capital letters refer to non normalized parameters and lower case letters to normalized values. τ and its Ψ did not need normalization.
A s = 0.76 cm 2 and e approximately 0.3 mm, which gives r i ≈ 0.26 R i and r e ≈ 0.26 R e . e non normalized and normalized parameters are of similar magnitude. We will continue to use R i and R e for the rest of the work.
Equivalent circuit parameters of the distributed circuit model (single-DCE, gure 1) were estimated by using the method of non-linear least squares curve-tting program using Microso Excel.

Statistical analysis
Relations between the leaves properties and electrical parameters were studied. e Pearson correlation coe cient (r) was calculated and the t-test was used to estimate the signi cance of correlation. Statistical analyses were carried out with SPSS so ware.

( )
i r Z f Z = in complex plane between 40 Hz to 100 kHz for three young leaves from the same branch of the tree number 4 are presented (Figure 2). e impedance spectrum of every leaf had a single arc in the form of parabola where the top corresponds to the frequency value characteristic of material f c and the intersection of the parabola with the x axis gives R ∞ and the Re. e radius of the arc increased with increase in leaf maturity.  All resistance and τ parameters increased during the study ( gure 4). e R i was 60Ω at the beginning of July (before bark stripping) and rose up to 105Ω in August ( gure 4a). e extracellular resistances typically were the lowest (35kΩ) at the bark stripping date and then increased ( gure 4b), τ inreased slightly from 1.33ms to 1.81ms a er bark-stripping ( gure 4c); i.e. the characteristic frequency fc decreased from 120 to 88Hz. Comparatively to intracellular resistance we noticed a slight increase R e already seven days a er bark-stripping, however, 14 days a er stripping, there was a strongly di erence between the variations of intra and extracellular resistance (98% and 75% respectively) ( gure 4a and 4b). e exponent Ψ increased during experiments (from 0.63 to 0.78) ( gure 4d). Leaves MC was the highest for all treatements at beginning of the study and then tapered o ( gure 4e).
CD had a signi cant e ect on all EIS parameters and MC (P<0.01). ere was good correlation between MC and all EIS parameters except for Ψ (Table 2). Our results also demonstrated the importance of sampling date on EIS parameters (Table 2) and ( gure 5), espacially on the intracellular resistance gure 5a. young leaf by impedance analysis and dry mater content in the leaf. T 1 and T 2 served as reference trees, T 3 and T 4 , bark-stripped for the rst time with CD respectively 1.1 and 1.22, T 5 and T 6 bark-stripped one time with CD respectively 1.64 and 1.74.
Bars indicated standard errors.

DISCUSSION
It is shown that the theoretical values are in good agreement with the experimental ones. e occurrence of one arc in a typical impedance spectrum for the leaves and the satisfactory t of the single DCE model strongly indicate the suitability of the model in the study of leaves. is result is in agreement with the previous studies of olives leaves (Mancuso 1999), Rhododendron leaves (Väinölä and Repo 2000) or of silver birch leaves (Repo et al. 2004).
Cellular responses of cork oak young leaves to stress caused by bark stripping were detected by EIS, and the di erent EIS parameters were sensitive to di erent factors. e e ects of bark stripping were more evident 21 days a er bark stripping. e increase of intra and extracellular resistances could be in relation with water losses a er stripping, actually as demonstrated by Repo et al. (2007) the drying of the cells may be enhanced by a reduction in the transpiration stream a er leaf abscission, subsequently leading to an increase in both resistances parameters. e increase of intracellular resistance during the summer additionally with the stripping e ect could indicate a decrease in electrolytic content and the increase in cell sugar. In fact, as mentioned by Mancuso (1999), the decrease of intracellular resistance in olives leaves during winter was a result of an increase in electrolytic content and a decrease in cell sugar concentration. e di erence between intra-and extracellular increases (98% and 75% respectively) 14 days a er sttripping can partly be explained by forces binding water molecules in the cell. In fact, extracellular water (water in intercellular spaces between the cells) is strongly bound by imbibition force whereas intracellular water is retained by osmotic type force. e decrease in all resistances 21 day a er bark stripping can be explained by plant reaction to balance water de cit by stomatal regulation (openning and closing the stomates); water losses are reduced and the plant tends to nd its moisture. e increase of relaxation time ( ) with higher stripping coe cient meant that the characteristique frequency value (f c ) moved toward lower frequencies.
Changes in relaxation time suggest a change in cell membrane (cf. Repo et al. 2004). Accordingly, we concluded that excessive stripping may a ect the frequency response of cell membranes. e distribution coe cient has the mean for about 0.71 in cork oak leaves versus 0.66 in silver birch leaves (Repo et al. 2004) and 0.61 in tea leaves (Mizukami 2007). In the dielectric materials which consists of di erent types of molecules, e.g. biological samples, multiples sources may contribute to this parameter, including biological variability, cell orientation anisotropy, intracellular heterogeneity, multiple component structures and the existence of a range of di erent relaxation processes (Repo et al. 2004). In isotropic samples with even sizes, e.g. water molecules, the impedance arc is almost a semicircle and the distributed coe cient is close to 1 (Torgovnikov 1993). e value of ψ in cork oak leaves suggest that several interfaces exist that fact also a ect the relaxation time.
Moisture content and all electrical parameter curves aim to an asymptote 21 days a er bark stripping. ey are two explanations for this phenomenon: either the reaction of the tree to balance its hydric gap by progressive decrease of transpiration at the level of the leaves (Correia et al.1992), or as referred to Natividade (1950), by the formation of a new phellogen layer with immediate cork regeneration function which, due to cork's impermeability, limits the water losses by evaporation. e time of sampling was important for EIS parameters of cork oak leaves especially on the 7 th of August i.e. 30 days a er bark stripping (table 2). R i is the electrical parameter which had higher Pearson coe cient with MC, CD and ND.

CONCLUSION
In conclusion, this was the rst time that EIS was applied to study the e ect of bark stripping on the dehydration stress of the Quercus suber L. leaves. Impedance parameters were signi cantly sensible to physiological variations in the leaves that were associated with stresses caused by bark stripping. Intra and extracellular resistance increase as moisture content decreases and increase as bark stripping coe cient increase. According to this technique, the trees seemed to balance their hydric gap 21 days a er stripping. However, bark stripping occurred during the summer when trees were already subjected to other environmental stresses; consequently, there were a great number of variables over which we had no control. Further research is needed to evaluate the so called speed of balance hydric gap with an aim to de ne speci c bark stripping coe cient for each tree depending on the tree characteristics.
APPENDIX 1 e single DCE is composed of one distributed element (DCE) in series witk a resistor (R∞). e DCE is composed of a parallel arrangement of a resistor R and a constant phase element (CPE) ( gure 1). e impedance of the constant phase element is (Macdonald 1987(Macdonald , 1990(Macdonald , 1995:

APPENDIX 2
Working expression for equivalent circuit modeling according to the expressions 7 and 8

LISTE OF ABBREVIATIONS
CD : bark-stripping coe cient, is calculated as the ratio of the maximum length by the trunk perimeter at breast height ND : number of bark-stripping, is the number of time that the tree was stripped MC : moisture content, % R i : intracellular resistance, Ωm R e : extracellular resistance, Ωm r i : intracellular speci c resistance, Ωm (obtained by normalization of Ri ) r e : extracellular speci c resistance, Ωm (obtained by normalization of Re ) e : thickness of the leaf, m As : cross-sectional area, m2 α : angle between x axis and the line passing by the origin of the axes and the center of the arc in the form of parabola, deg fc : characteristic frequency value, Hz ω : the angular frequency, radians per second τ : relaxation time, s ψ : the distributed coe cient of relaxation time i : the imaginary unit EIS : electrical impedance spectroscopy DCE : distributed circuit element CPE : Constant Phase Element R∞ : represents the resistance at height frequency