Polar Biology

, Volume 26, Issue 8, pp 525–529

Response of rhizosphere microbial communities associated with Antarctic hairgrass ( Deschampsia antarctica) to UV radiation


  • L. M. Avery
    • School of Life and Environmental SciencesUniversity of Nottingham
    • School of Agricultural and Forest SciencesUniversity of Wales
  • R. I. Lewis Smith
    • British Antarctic SurveyNatural Environment Research Council
    • School of Life and Environmental SciencesUniversity of Nottingham
Original Paper

DOI: 10.1007/s00300-003-0515-y

Cite this article as:
Avery, L.M., Lewis Smith, R.I. & West, H.M. Polar Biol (2003) 26: 525. doi:10.1007/s00300-003-0515-y


The response of rhizosphere microbial communities associated with natural populations of Deschampsia antarctica growing on Léonie Island (67°36′S, 68°21′W, Antarctic Peninsula) to UV radiation was investigated. UV radiation was controlled in the field using Perspex VA screens (UV-B opaque) which transmit little radiation below 380 nm but allow penetration of approximately 92% of radiation above 400 nm, and Perspex OXO2 screens (UV-B transparent) which transmit approximately 70% of radiation at 280 nm, rising to 90% at 300 nm and above. Reducing ambient UV radiation altered the phenotypic profile of the rhizosphere microbial community. This alteration was expressed as enhanced carbohydrate and carboxylic acid utilisation by the rhizosphere micro-organisms. It is hypothesised that ambient levels of UV radiation indirectly affect rhizosphere micro-organisms by influencing the quality or quantity of root exudates.


Since the mid-1970s stratospheric ozone depletion has occurred as a result of increased concentrations of chlorofluorocarbons (CFCs) and other halon gases in the upper atmosphere (Kerr 1988; Russell et al. 1996). Antarctica now undergoes an annual spring ozone depletion event during which losses relative to the 1970s are estimated to be 50%, with removal of almost all of the ozone at altitudes of 15–20 km during a 6-week period (Pyle 1997; Madronich 1999). In September 2000 the Antarctic ozone hole covered ca. 28.4 million km2 and ca. 8 million km2 in 2002 (Shanklin 2003). Despite the relatively smaller coverage, intensity reached its highest values at the British Antarctic Survey's (BAS) Rothera research station on Adelaide Island, south-western Antarctic Peninsula, between mid-September and early October 2002, since site records began in 1997 (Shanklin 2003). Recent predictions suggest that ozone depletion due to anthropogenic activity is currently close to its maximum and some recovery should occur over the next few decades. However, enhanced levels of UV-B (280–320 nm) reaching the planet's surface are expected to persist well into the 21st century (Madronich et al. 1995). Deschampsia antarctica Desv. (Poaceae) and Colobanthus quitensis (Kunth) Bartl. (Caryophyllaceae) are the only native vascular plants of Antarctica. They are widespread, usually growing together, in low altitude coastal sites throughout the maritime Antarctic where climate, habitat and edaphic conditions provide oases for the growth and development of these plants in an otherwise hostile environment (Lewis Smith 1997, 2003). The population studied in our experiments is close to the southern limit of the species and, in summer, is exposed to a much longer duration of sunshine and higher UV-B levels than populations growing in the northern maritime Antarctic (e.g. South Shetland and South Orkney Islands). Several recent investigations have addressed the impact of UV-B on D. antarctica. Day et al. (1999), using a field cloche system, showed that reproductive development was more rapid under ambient UV, and the authors suggest this might be a stress response in which reproductive growth takes priority over vegetative growth. Rozema et al. (2001), in a controlled environment experiment using D. antarctica collected from Léonie Island, near Rothera station, demonstrated that UV-B reduced shoot length but increased shoot number and leaf thickness. The content of UV-B absorbing compounds was not enhanced with increasing UV-B in their experiment, which is in agreement with the findings of Day et al. (1999) and Lud et al. (2001). Despite UV-B mediated differences in carbon allocation, photosynthesis in D. antarctica appeared to be unaffected (Huiskes et al. 2001; Lud et al. 2001; Rozema et al. 2001). These findings are significant, since many studies of other plant systems indicate that UV-B modifies not only leaf and shoot architecture, but also production of secondary metabolites (e.g. Rozema et al. 1997 and references therein).

Until the present study, nothing was known about the responses of any soil microbial communities in Antarctica to UV-B radiation. The most diverse and active microbial communities occur in the loam-like soil formed beneath some of the more extensive closed stands of D. antarctica, and these rhizosphere micro-organisms are important in supplying plant nutrients; similarly, microbial growth and activity are dependent on root exudates and rhizodeposits. Thus, it is to be expected that any plant-mediated change in response to abiotic factors will inevitably affect the rhizosphere microbial community.

Our study is the first field-based experiment in Antarctica to investigate plant-mediated responses of rhizosphere micro-organisms to solar UV radiation. We report the results of two experiments in which UV radiation levels were manipulated using UV-B absorbing and UV-B transparent Perspex screens over naturally growing swards of D. antarctica. We hypothesised that reducing ambient levels of UV-B would cause a shift in the phenotypic profile of rhizosphere micro-organisms towards greater metabolic diversity, because carbon, which may otherwise be allocated to shoots in response to UV-B stress, would instead be directed to the roots.

Materials and methods

Study site

The field experiments were established by R.I.L.S. on Léonie Island (67°36′S, 68°21′W), in northern Marguerite Bay, about 15 km west of the Antarctic Peninsula and 7.5 km west of the BAS research station at Rothera Point on south-east Adelaide Island. Léonie Island is in the lee of the much larger Adelaide Island, and experiments were conducted on dense swards of D. antarctica growing on the sheltered north side of the island, which is free of snow during the austral summer. The principal associated species were C. quitensis and the moss Sanionia uncinata.

Field manipulations

Two replicate Perspex screens for each UV treatment (ambient and sub-ambient) were deployed during February 1995 and a further two in January 1998 to give a long-term (4 year) and a short-term (1 year) experiment, respectively. Thus, a total of eight screens were deployed. The north-facing screens were randomly placed over closed swards of D. antarctica and were located approximately 1–2 m apart within each experiment; the short- and long-term experiments were about 5–10 m apart.

Screens consisted of aluminium frames (50×50 cm) covered with one of two types of 3-mm-thick transparent Perspex filter materials to effect UV treatments. Perspex VA cast acrylic sheet (Du Pont Polyesters Group, Middlesborough, UK) was used to exclude both UV-B and UV-A. Perspex VA transmits virtually no radiation below 380 nm but does transmit approximately 92% of radiation above 400 nm. Perspex OXO2 extruded acrylic sheet (Du Pont Polyesters Group) was used to cover frames through which UV-B and UV-A penetrated. Perspex OXO2 transmits approximately 80% of UV-B and 90% of UV-A. Transmission of radiation at 280 nm is approximately 70% rising to 90% at 300 nm and above. The open-sided screens were 15 cm high at the uphill (south) face, sloping to 10 cm at the downhill (north) face to allow snow melt and rain run off. The tops of the screens consisted of two slightly overlapping Perspex sheets, with a 1 cm gap within the overlap for ventilation. Temperatures within the screens were not measured throughout the duration of the experiments for logistical reasons. However, Lud et al. (2003), used similar open-sided screens in their study of the influence of UV-B on photosynthetic performance of S. uncinata on Léonie Island. They recorded temperature using Tiny Talk II mini-dataloggers (Gemini data loggers, Chichester, UK) in the moss carpet by placing thermocouples at the surface of the vegetation for 17 days in January and February 1998, and for 21 days in January and February 1999. Temperatures ranged from −1.5 °C to 22.8 °C in 1998 and from −1.8 °C to 24.9 °C in 1999 within the control (unscreened) moss carpet. Temperature under the filter screens was 0.6 °C and 0.7 °C higher in 1998 (night and day, respectively) and 0.1 °C and 6.3 °C higher in 1999 (night and day, respectively) compared with uncovered surrounding vegetation. It is likely that the D. antarctica sward was subjected to a similar temperature range.


Three cores of soil/rhizosphere material were collected from beneath each screen during the austral summer on 9 February 1999. Each core was approximately 10 cm in depth and diameter and the grass remained intact on the surface. Cores were frozen at −20 °C within 2–3 h of collection. They remained frozen for approximately 10 weeks until they were returned to Nottingham. Prior to analysis, cores were thawed slowly in a constant temperature room (4 °C) for 72 h.

Culture of microbes from rhizosphere soil

Each core was manually homogenised (in a sterile bag). Culturable bacterial populations were extracted by performing standard 10-fold serial dilutions, plating onto duplicate Tryptone Soya agar (Oxoid) and incubating at 10 °C and 20 °C for 48 h. A representative number of cultured bacterial colonies were selected randomly from each plate giving a total of 60 isolates per filter screen. These isolates were individually maintained in axenic culture and were Gram stained and assessed for their ability to hydrolyse starch, urea and protein using a modified method of Klironomos and Allen (1995). Urea and starch plates were incubated at 20 °C for 24 and 48 h respectively. Nutrient gelatin plates were incubated at 20 °C for up to 120 h and examined daily for liquefaction reactions. These tests and isolations were carried out for both the long- and short-term experiments.

Three commercially prepared microtitre plates (Biolog GN, Biolog California, USA) per filter screen were inoculated with rhizosphere soil dilutions from each soil sample to provide a community-level physiological profile of the rhizosphere micro-organisms. A standardised inoculum equating to 5×102 colony forming units (CFU) ml-1 (based on the results of previous plate count data) was used to inoculate an individual Biolog GN plate for each rhizosphere sample, thus giving triplicate plates for each filter screen. Community-level physiological profiles were determined on rhizosphere soil samples originating from screens established in 1995 (long-term experiment). Biolog plates were incubated at 20 °C for 5 days and were read daily by scoring each well on a 2-point scale (distinguishable change or no change) for colour development (Ellis et al. 1995). The 95 substrates within the wells were grouped into compound types, the values in parenthesis refer to the number of different substrates corresponding to the specific substrate grouping (Garland and Mills 1991): alcohols (2), amides (3), amines (3), amino acids (20), aromatic chemicals (4), brominated chemicals (1), carbohydrates (28), carboxylic acids (24), esters (2), phosphorylated compounds (3) and polymers (5). The proportion of wells of each substrate type showing utilisation (colour development) and the proportion of the total number of substrates that were utilised at each time point were used in data analysis.

Statistical analyses

A repeated measures one-way analysis of variance (ANOVA) was carried out on data for protein hydrolysis and Biolog to test for the main effect of UV-B radiation at each reading time. A one-way ANOVA was performed on data for starch and urea hydrolysis, Gram stain data and culturable bacterial counts (log10 transformed). The least significant difference (LSD) was determined to test the significance of the differences between group means. All data were analysed using GenStat 6.1 (Lawes Agricultural Trust, IACR-Rothamsted, UK). Data sets from the long-term and short-term experiments were analysed separately.

Results and discussion

At Rothera station spectral irradiance data were obtained from a Bentham spectroradiometer (Bentham DM 150, Bentham Instruments, Reading, UK) which measures the spectral global irradiance between 280 and 600 nm. These data were used to calculate UV-B and UV-A levels (defined for these parameters as 280–315 and 315–400 nm, respectively) for the part of the experimental period for which they are available, and are expressed as general plant damage weighted values and actual daily dose values of UV-B radiation reaching Rothera Point. Recordings began during February 1997. Photosynthetically active radiation (PAR) was also estimated from the readings using a model which defines PAR as 400–700 nm (Fig. 1). Spectral global irradiance is not measured during the dark winter months.
Fig. 1.

Daily dose of UV-B radiation (heavy lines) and general plant damage weighted daily dose of UV-B radiation (fine lines) measured using the Bentham spectroradiometer situated on the roof of the Bonner Laboratory at Rothera research station (67°34′S, 68°08′W) during the period 3 February 1997 to 28 February 1999. In the calculation of these data UV-B was defined as 280–315 nm, UV-A as 315–400 nm and photosynthetically active radiation (PAR) was estimated from the spectral irradiance measurements (280–600 nm) using a model to give PAR as 400–700 nm. Data provided by H. Peat (British Antarctic Survey)

A significant UV-related change in utilisation patterns of carbohydrates and carboxylic acids by plant-associated microbial communities was observed (Table 1). Communities originating from the rhizosphere of D. antarctica grown in the presence of ambient levels of UV radiation, utilised proportionally fewer carbohydrates than those originating from the rhizosphere of plants grown at sub-ambient levels of UV radiation. This pattern was consistent at each time period measured. Carboxylic acids were utilised to the same extent by each community type after 72 h, but utilisation by microbes associated with plants grown at sub-ambient UV was significantly greater 48 h after inoculation. The consistent trend of reduced carbohydrate usage by rhizosphere microbes when plant associates had been grown with ambient UV indicates that the physiological profiles of the two soil microbial communities were different. Since Biolog profiles may reflect carbon source availability (Grayston et al. 2001), our data suggest that ambient UV radiation lessened availability of carbon sources within the rhizosphere. Johnson et al. (2002) demonstrated in an experiment on subarctic heath at Abisko, Sweden, that increased UV-B in combination with enhanced CO2 led to reduced microbial utilisation of C sources in Biolog plates. They concluded that observed shifts in the metabolic profile of the soil micro-organisms resulted from a treatment-related change in dominant bacterial species within the wells. Their conclusion is probably also pertinent to the present study. It is possible that UV-mediated changes in plant chemistry induced selective pressure on the D. antarctica rhizosphere microbial community through qualitative and quantitative alterations in rhizodeposition. Indeed, Xiong and Day (2001) reported decreased concentrations of UV-B absorbing compounds and chlorophyll content in D. antarctica under reduced UV-B, in tandem with increased leaf production and a tendency towards enhanced root biomass. UV-B induced alterations in plant chemistry are well reported (e.g. Rozema et al. 1997) and it is known that UV-B radiation stimulates accumulation of UV-absorbing polyphenols by affecting phenylpropanoid and flavonoid biosynthesis gene expression (Beggs and Wellmann 1994). Distribution of epidermal flavonoids increased 2-fold in Secale cereale when exposed to UV-B (Anhalt and Weissenböck 1992), whereas phenolic molecules in the mesophyll were unaffected. More recently, Newsham et al. (2001) reported small (3–6%) increases in soluble carbohydrates (arabinose and glucose) in leaves of Quercus robur in response to elevated UV radiation. Gehrke et al. (1995) had previously demonstrated that the proportions of soluble carbohydrates and tannins in leaf litter originating from Vaccinium myrtillus grown under enhanced UV-B were greater than in corresponding litter from unexposed plants. A 37% increase in holocellulose concentration in Triticum aestivum leaves after exposure to UV radiation (simulating a 25% ozone reduction) was reported by Yue et al. (1998), and Gwynn-Jones (2001) demonstrated increases in respiration accompanied by a reduction in root soluble carbohydrate content and soluble sugar:starch ratio in shoots of Calamagrostis purpurea subjected to enhanced UV-B in an outdoor facility at Abisko. Gwynn-Jones (2001) suggested that long-term damage caused by UV-B may, in effect, be resource depletion of carbohydrates and therefore an inability to support protective and repair mechanisms.
Table 1.

Community level carbon utilisation and culturable bacteria associated with the rhizosphere of Deschampsia antarctica. Natural swards were screened for 4 years under UV-B transparent or UV-B opaque filters


Reading time (h)

Mean % total substrates oxidised

Mean % carbohydrate substrates oxidised

Mean % carboxylic acids oxidised

Mean % amino acids oxidised

Bacteria cultured at 10°C (log10 cfuc g–1 dry wt. soil)

Bacteria cultured at 20°C (log10 cfuc g–1 dry wt. soil)

Near ambient


























































Analyses of variance















UV-B × time







a Least significant difference (LSD) calculated at 5% level for UV-B as a single factor

b Calculated at 5% level for UV-B × time interaction

c Colony forming units

In our study, the proportion of the 95 substrates utilised by communities originating from screens allowing penetration of UV radiation was consistently lower at each reading time compared with the proportion oxidised by communities from under filters which screened out UV. However, this trend was not significant and probably reflects the fact that utilisation of other substrate groupings, such as amino acids, was not influenced by UV and that other compounds (e.g. phosphorylated chemicals and amides) were only oxidised 72 h and 96 h after inoculation, respectively. All other groups of chemicals were utilised to differing extents within 48 h of inoculation. We are confident of the validity of the data since the scoring system used was robust as it accounted only for definite changes in utilisation.

Reducing ambient levels of UV radiation in either the long- or short-term experiments, did not affect starch, urea or protein hydrolysis by rhizosphere bacterial isolates. Numbers of culturable bacteria and the proportion of Gram positive and Gram negative bacteria were also unaffected by UV. This is in contrast to the findings of Klironomos and Allen (1995) who reported UV-B mediated increases in total bacterial numbers, the proportion of Gram negative bacteria and starch hydrolysing bacteria as well as decreases in urea- and protein-hydrolysers. These authors state that their findings are suggestive of an increase in rhizodeposition and a shift in quality of exudates in favour of compounds that were less nitrogen-rich. Our data do not substantiate these findings, although they do suggest that root exudate quality is altered by UV in terms of carbon substrates.

The filter-screen system used in this investigation has been subject to some criticism (e.g. Huiskes et al. 2001) because the microclimate within screens differs from that outside. Although this should be taken into consideration, the system has the advantage of maintaining realistic PAR levels and is easily maintained under the inclement environmental conditions of Antarctica. The aim of this current investigation was to determine whether the rhizosphere microbial community is influenced by UV radiation. Since we were comparing screen with screen, any screen-related alterations in, for example, moisture availability should be similar throughout. The main constraint within this investigation is the low level of replication (two screens per UV treatment and three pseudoreplicates per screen), although the impact of UV as a single factor on carbohydrate utilisation is clear. In summary, we have demonstrated that reducing ambient UV radiation altered the metabolic profile of rhizosphere microbial communities associated with natural populations of D. antarctica. Such changes within the rhizosphere have probably also occurred naturally within the terrestrial ecosystem, since the current UV-B levels to which the vegetation is exposed during the Antarctic spring and early summer have increased considerably over the past quarter of a century. By reducing the current ambient UV-B levels in the experiments described here, the rhizosphere may be regarded to have reverted to a condition approaching that at the commencement of stratospheric ozone depletion in the mid-1970s.


We are grateful to Dr. P. Convey and A. Rossaak for collecting the field samples, Dr. H.J. Peat for providing the Bentham Spectroradiometer data, and Dr. S. Harangozo for providing meteorological data. Drs K.K. Newsham and A.H.L. Huiskes supplied useful comments on the manuscript for which we are grateful. Lisa Avery was in receipt of a University of Nottingham studentship whilst this work was undertaken.

Copyright information

© Springer-Verlag 2003