Interactive signal transfer between host and pathogen during successful infection of barley leaves by Blumeria graminis and Bipolaris sorokiniana

Using ion-selective microprobes, interactive signalling between barley and Blumeria graminis or Bipolaris sorokiniana has been investigated. The question was raised whether a biotrophically growing fungus manipulates the electrical driving forces (membrane potential, transmembrane pH), required for H+ cotransport of energy-rich compounds. Electrodes were positioned in the substomatal cavity of open stomata or on the leaf surface, and pH was measured continuously up to several days during fungal development. We demonstrate that surface and apoplastic fluids are electrically coupled and respond in a similar manner to stimuli. Apoplastic pH, monitored from the moment of inoculation with conidia, reveals several phases: 2-4h after inoculation of the barley leaf with either fungus, the host displays rapid transient responses after its first contact with the fungal cell wall; apoplastic pH and pCa increases, cytoplasmic pH and pCa decreases. About 1 day after inoculation, the apoplastic pH increases by up to 2 pH units, which is thought to reflect a resistance response against the intruder. Whereas barley leaf cells possess a membrane potential of -152+/-5 mV, hyphae of B. graminis yield -251+/-8 mV, indicative of a substantial driving force advantage for the fungus. Although the resting membrane potential of barley remains constant during the first days after inoculation, leaves infected with B. sorokiniana get confronted with an energy problem, indicated by a retarded repolarization following a "light-off" stimulus. Five days after inoculation, apoplastic pH has increased to 5.97+/-0.47 (n=11) and does no longer respond to "light-off" when measured within lesions. In contrast, it stays at near normal values outside the lesions and responds to "light-off". It is concluded that biotrophically growing fungi do not manipulate the cotransport driving forces since (i) any change in apoplastic pH would be experienced by both partners; (ii) the resting membrane potential is not changed. It is suggested that measured pH changes reflect defence responses of the host against the fungus rather than fungal action to increase compatibility.     

Interactions of Neotyphodium gansuense, Achnatherum inebrians, and plant-pathogenic fungi

Interactions of Neotyphodium gansuense, Achnatherum inebrians, and nine fungal pathogens were studied by tests of inhibition of four fungal pathogens by Neotyphodium endophytes in vitro and by inoculation of nine fungal pathogens on detached leaves of endophyte-infected (E+) and endophyte-free (E−) plants. Compared with the controls, most isolates of N. gansuense significantly inhibited the growth in vitro of, in decreasing order of inhibition, Bipolaris sorokiniana, Curvularia lunata, Fusarium acuminatum, and Alternaria alternata. Inhibition zones appeared between pathogens and some isolates of N. gansuense. Some isolates of N. gansuense significantly inhibited sporulation of B. sorokiniana, A. alternata, and C. lunata. However, there was no significant inhibition of F. acuminatum and a few isolates significantly increased sporulation. The leaf inoculation trial indicated that almost all fungal pathogens were able to cause lesions on detached leaves regardless of endophyte status. Both the number and size of disease lesions on E+ A. inebrians leaves caused by A. alternata, F. chlamydosporum, F. oxysporum, and F. solani were reduced compared with those on E− leaves. Only lesion numbers (not size) of Ascochyta leptospora leaf spots were significantly reduced on E+ leaves compared with E− leaves. Conversely, only the length of Ascochyta leptospora leaf spots were significantly smaller on E+ leaves than on E− leaves; numbers of lesions were not significantly affected. C. lunata was strongly pathogenic to both E+ and E− leaves and numerous lesions developed and merged into patches, the leaf surface was covered and the leaf rotted away.     

Increased Levels of Cytokinins in Barley Leaves Having the Systemic Acquired Resistance to Bipolaris sorokiniana (Sacc.) Shoemaker

The role of endogenously induced higher level of cytokinins and exogenously applied kinetin in relation to the development of barley leaf spot caused by Bipolaris sorokiniana (Sacc.) Shoemaker (syn. Helminthosporium sativum Pammel, King and Bakke) was studied. Spraying barley leaves with kinetin suppressed the number and the size of necrotic spots caused by the fungus. Inoculation of the lower leaves of barley by a spore suspension of the fungus B. sorokiniana induced resistance on the upper leaves against a subsequent challenge inoculation by the same pathogen 10 days later. An increase in the level of cytokinins was observed in these resistant leaves. Elevated levels of cytokinins may cause a type of juvenility in leaf tissues. The juvenile state could be in a causal relationship with the suppression of necrotic spots caused by the fungus.     

Ultrastructural and Histochemical Studies on Mildew of Barley (Erysiphe graminis DC. f. sp. hordei Marchal)

Plants of the mildew susceptible barley cultivar Peruvian and the adult plant resistant cultivar Osiris were inoculated with Erysiphe graminis f. sp. hordei at the first and fifth leaf stages. Samples taken at 32 hr after inoculation were examined by electron microscopy to compare papillae associated either with penetration failure or with successful penetration of the fungus into the epidermal cell and haustorium formation. Four types of papillae with ultrastructural differences could, be classified. Although their definite association with fungal ingress or failure is not possible, our data suggest that papillae with larger, more compacted and amorphous or globular structures may be more effective as penetration barriers than others, with more or less uniform distribution of irregular, smaller electrondense structures.     

Infection process of Gaeumannomyces graminis var. graminis on the roots and culms of rice

Crown sheath rot, caused by the ascomycete Gaeumannomyces graminis var. graminis that infects the root and the base of the culm of rice, causes early grains maturation, tiller death and reduced yield. As a paucity of information exists in the literature on the rice‐G. graminis var. graminis interaction at the microscopic level, this study aimed to gain novel insights into the infection process of this pathogen in the root and culm of rice using both light and scanning electron microscopy. In the roots, the fungus initially colonized the epidermal, exodermal and sclerenchyma cells. At 15 days after inoculation (dai), fungal hyphae colonized the cortex and clusters of perithecia were observed in the roots. At 20 dai, the fungus reached the central cylinder, and an intense fungal colonization at the base of the culm was observed that resulted in the formation of a mycelial mat on both adaxial and abaxial surfaces of the leaf sheaths. At 25 dai, fungal growth was noticed in the parenchyma cells, vascular bundles and airspaces. Perithecia emerged through the base of prophyllum and from the first leaf sheath at 30 dai. The results of this study provide new insights into the infection process of G. graminis var. graminis in rice and may help to find better control measures in reducing crown sheath rot development.     

Pathogen-Induced Changes in the Antioxidant Status of the Apoplast in Barley Leaves

Leaves of two barley (Hordeum vulgare L.) isolines, Alg-R, which has the dominant Mla1 allele conferring hypersensitive race-specific resistance to avirulent races of Blumeria graminis, and Alg-S, which has the recessive mla1 allele for susceptibility to attack, were inoculated with B. graminis f. sp. hordei. Total leaf and apoplastic antioxidants were measured 24 h after inoculation when maximum numbers of attacked cells showed hypersensitive death in Alg-R. Cytoplasmic contamination of the apoplastic extracts, judged by the marker enzyme glucose-6-phosphate dehydrogenase, was very low (less than 2%) even in inoculated plants. Dehydroascorbate, glutathione, superoxide dismutase, catalase, ascorbate peroxidase, glutathione reductase, monodehydroascorbate reductase, and dehydroascorbate reductase were present in the apoplast. Inoculation had no effect on the total foliar ascorbate pool size or the redox state. The glutathione content of Alg-S leaves and apoplast decreased, whereas that of Alg-R leaves and apoplast increased after pathogen attack, but the redox state was unchanged in both cases. Large increases in foliar catalase activity were observed in Alg-S but not in Alg-R leaves. Pathogen-induced increases in the apoplastic antioxidant enzyme activities were observed. We conclude that sustained oxidation does not occur and that differential strategies of antioxidant response in Alg-S and Alg-R may contribute to pathogen sensitivity.     

Ribulose Bisphosphate Carboxylase Protein and Enzymes of CO2 Assimilation in Barley Infected by Powdery Mildew (Erysiphe graminis hordei)

Effects were examined of barley powdery mildew (Erysiphe graminis f. sp. hordei) on the major leaf protein, ribulose-l,5-bisphosphate carboxylase (RuBPCase), and other enzymes of CO₂ assimilation, phosphoenolpyruvate carboxylase (PEPcase) and malic enzymes, and enzymes associated with RuBPCase in the reductive pentose phosphate pathway. Activity of RuBPCase per unit fresh weight of leaf was inhibited by infection from the first sample, 3 days after inoculation, to the last sample, 24 days after inoculation, when healthy control leaves were visibly senescing. The inhibition occurred because the amount of RuBPCase protein (measured specifically by an immunological technique) was reduced from 6 days after inoculation and because activity per unit protein declined from 3 days until 21 days after inoculation, at which time there was little protein remaining. Activity of PEPcase per unit fresh weight of leaf was initially stimulated by infection but, thereafter, it was inhibited. Inhibition, like that also affecting malic enzymes (NAD) and (NADP), 3-phosphoglycerate kinase, and glyceraldehyde-3-phosphate dehydrogenases (NAD) and (NADP), was associated with a decline in amounts of total soluble minus RuBP Case protein per unit fresh weight of leaf. Reduced amounts of leaf protein may be associated with reduced nitrate uptake by roots and fungal demand for nitrogen in mildewed plants.     

Classical biological control of Aphis gossypii (Homoptera: Aphididae) with Neozygites fresenii (Entomophthorales: Neozygitaceae) in California cotton

In August 1994 and 1995 classical biological control releases were made in cotton in the San Joaquin Valley, California, with an Arkansas strain of the entomopathogenic fungus, Neozygites fresenii, a pathogen of the cotton aphid, Aphis gossypii. Pre-release samples in both years indicated that N. fresenii was not naturally present in A. gossypii populations in the San Joaquin Valley. Two release methods were compared: dried N. fresenii-infected cotton aphid “cadavers” and chamber inoculation of A. gossypii. Both methods were successful in introducing N. fresenii to cotton aphids in California; however, higher prevalence of fungal infection resulted with the cadaver treatments. N. fresenii persisted and spread in the aphid population until early October 1994 and late September 1995. The highest mean percentage infection in the cadaver treatment in 1994 reached a level (14%) considered imminent for epizootics (12–15%). The use of predator exclusion cages resulted in higher N. fresenii prevalences.     

Phenylalanine Ammonia-lyase Activity in Barley After Infection with Bipolaris sorokiniana or Treatment with its Purified Xylanase

Phenylalanine ammonia-lyase (PAL) activity was determined from leaves and roots of two barley (Hordeum vulgare L.) cultivars after infection with a necrotrophic pathogen, Bipolaris sorokiniana (Sacc.) Shoem., and treatment with its purified xylanase. PAL activity increased in leaves of both cultivars 16 h after fungal inoculation but two phases, with activity peaks at 24–32 h and 40 h, were recorded only for the more resistant cultivar, Agneta. Attempts to use a PAL inhibitor, χ-amin, ooxyacetic acid, to increase susceptibility to B. sorokiniana in barley leaves were unsuccessful. Treatments of leaves with purified xylanase resulted in more rapid (4–12 h after injection), although reduced, induction of PAL compared with fungal injection. The higher the concentration of xylanase applied the earlier the activity peaks were detected. Fungal inoculation only slightly increased PAL activity in barley roots while xylanase treatment had no effect. The basal level of PAL was however much higher in roots than in leaves. In wheat, Triticum aestivum L. resistant to B. sorokiniana, the time-course of PAL induction after fungal infection and xylanase treatment resembled that for cv. Agneta, while in oats, Avena sativa L. (non-host) PAL activity did not change after the treatments. The results suggest that the second phase of PAL induction, associated only with responses of barley cv. Agneta and wheat, is linked with their resistance to B. sorokiniana infection. The possible role of xylanase as an elicitor of PAL is discussed.     

Seed germination and dormancy in the medicinal woodland herbs Collinsonia canadensis L. (Lamiaceae) and Dioscorea villosa L. (Dioscoreaceae)

We used a double germination phenology or “move-along” experiment (sensu Baskin and Baskin, 2003) to characterize seed dormancy in two medicinal woodland herbs, Collinsonia canadensis L. (Lamiaceae) and Dioscorea villosa L. (Dioscoreaceae). Imbibed seeds of both species were moved through the following two sequences of simulated thermoperiods: (a) 30/15 °C→20/10 °C→15/6 °C→5 °C→15/6 °C→20/10 °C→30/15 °C, and (b) 5 °C→15/6 °C→20/10 °C→30/15 °C→20/10 °C→15/6 °C→5 °C. In each sequence, seeds of both species germinated to high rates (>85%) at cool temperatures (15/6 and 20/10 °C) only if seeds were previously exposed to cold temperatures (5 °C). Seeds kept at four control thermoperiods (5, 15/6, 20/10, 30/15 °C) for 30 d showed little or no germination. Seeds of both species, therefore, have physiological dormancy that is broken by 12 weeks of cold (5 °C) stratification. Morphological studies indicated that embryos of C. canadensis have “investing” embryos at maturity (morphological dormancy absent), whereas embryos of D. villosa are undeveloped at maturity (morphological dormancy present). Because warm temperatures are required for embryo growth and cold stratification breaks physiological dormancy, D. villosa seeds have non-deep simple morphophysiological dormancy (MPD). Neither species afterripened in a 6-month dry storage treatment. Cold stratification treatments of 4 and 8 weeks alleviated dormancy in both species but C. canadensis seeds germinated at slower speeds and lower rates compared to seeds given 12 weeks of cold stratification. In their natural habitat, both species disperse seeds in mid- to late autumn and germinate in the spring after cold winter temperatures alleviate endogenous dormancy.     

The Phylogenetics of Mycotoxin and Sclerotium Production in Aspergillus flavus and Aspergillus oryzae

Aspergillus flavus is a common filamentous fungus that produces aflatoxins and presents a major threat to agriculture and human health. Previous phylogenetic studies of A. flavus have shown that it consists of two subgroups, called groups I and II, and morphological studies indicated that it consists of two morphological groups based on sclerotium size, called “S” and “L.” The industrially important non-aflatoxin-producing fungus A. oryzae is nested within group I. Three different gene regions, including part of a gene involved in aflatoxin biosynthesis (omt12), were sequenced in 33 S and L strains of A. flavus collected from various regions around the world, along with three isolates of A. oryzae and two isolates of A. parasiticus that were used as outgroups. The production of B and G aflatoxins and cyclopiazonic acid was analyzed in the A. flavus isolates, and each isolate was identified as “S” or “L” based on sclerotium size. Phylogenetic analysis of all three genes confirmed the inference that group I and group II represent a deep divergence within A. flavus. Most group I strains produced B aflatoxins to some degree, and none produced G aflatoxins. Four of six group II strains produced both B and G aflatoxins. All group II isolates were of the “S” sclerotium phenotype, whereas group I strains consisted of both “S” and “L” isolates. Based on the omt12 gene region, phylogenetic structure in sclerotium phenotype and aflatoxin production was evident within group I. Some non-aflatoxin-producing isolates of group I had an omt12 allele that was identical to that found in isolates of A. oryzae.     

Biological control of alligatorweed (Alternanthera philoxeroides) with a Fusarium sp.

Alligatorweed (Alternanthera philoxeroidesG.) has become a serious weed in different crops in China. A fungal pathogen was found in Chongqing and Sichuan Provinces and was identified as a species in the Fusarium genus. The fungus produced macroconidia and chlamydospores abundantly on potato sucrose agar (PSA) plates. The bestconidial production and germination and colonygrowth of Fusarium sp. were at 23–31°C and pH 6.7–7.0. Light period and flooding did not affect fungal growth and conidium formation. The herbicides, glyphosate and paraquat, inhibited the fungal development in vitro. The fungus did not affect seed germination and seedling growth of paddy rice, wheat, maize, oilseed rape and broad bean inlaboratory or greenhouse trials. Inoculum density and wetness duration influenced the efficiency of Fusarium sp. to control alligatorweed; a concentration of 1.0 × 10⁵ spores^–1 ml and 12 h of high humidity duration after inoculation produced goodinfections on the weed at 23°C in the laboratory. When the fungus was applied to alligatorweed grown in greenhouse and in the field, good biocontrol efficiency was obtained: the plants started to wilt after four to five (greenhouse) or six days (field), and were killed 9–10 (greenhouse) or 13–14 (field) days after spraying the fungal inoculum. This was similar to the control efficiency resulting from glyphosate treatment. Therefore, this Fusarium sp. appeared to be a good candidatefor further studies and a promising biocontrol agent to manage alligatorweed in some terrestrial and aquatic crops.     

Leaf scarring by the weevils Neochetina eichhorniae and N. bruchi enhances infection by the fungus Cercospora piaropi on waterhyacinth, Eichhornia crassipes

Additive or synergistic effects among introduced and native insect and plant pathogen agents are necessary to achieve biological control of waterhyacinth (Eichhornia crassipes), a globally damaging aquatic weed. In field plots, plants were infested with waterhyacinth weevils (Neoechetina bruchi and N. eichhorniae) and leaves were scarred by weevil feeding. Subsequent infection by the fungal pathogen Cercospora piaropi caused necrotic lesions to form on leaves. Necrosis development was 7.5- and 10.5-fold greater in plots augmented with both weevils and C. piaropi and weevils alone, respectively, than in plots receiving only C. piaropi. Twenty-four days after weevil infestation, the percentage of laminar area covered by lesions on third-youngest and oldest live leaves was elevated 2.3–2.5-fold in plots augmented with weevils. Scar density and necrosis coverage on young leaf laminae were positively correlated, even though antipathogenic soluble peroxidases were elevated 3-fold in plots augmented with weevils alone or weevils and C. piaropi. Combined weevil and fungal augmentation decreased shoot densities and leaves per plant. In a no-choice bioassay, weevil feeding on oldest but not young leaves was reduced 44 two weeks after C. piaropi inoculation. Protein content and peroxidase activities were elevated 2–6-fold in oldest leaves three weeks after inoculation. Augmentation with both waterhyacinth weevils and C. piaropi led to the development of an additive biological control impact, mediated by one or more direct interactions between these agents, and not plant quality effects.     

Antioxidant defences of the apoplast

Summary The apoplast of barley and oat leaves contained superoxide dismutase (SOD), catalase, ascorbate peroxidase, dehydroascorbate reductase, monodehydroascorbate reductase, and glutathione reductase activities. The activities of these enzymes in the apoplastic extracts were greatly modified 24 h after inoculation with the biotrophic fungal pathogenBlumeria graminis. The quantum efficiency of photosystem II, which is related to photosynthetic electron transport flux, was comparable in inoculated and healthy leaves during this period. Apoplastic soluble acid invertase activity was also modified in inoculated leaves. Inoculation-dependent increases in apoplastic SOD activity were observed in all lines. Major bands of SOD activity, observed in apoplastic protein extracts by activity staining of gels following isoelectric focusing, were similar to those observed in whole leaves but two additional minor bands were found in the apoplastic fraction. The apoplastic extracts contained substantial amounts of dehydroascorbate (DHA) but little or no glutathione (GSH). Biotic stress decreased apoplastic ascorbate and DHA but increased apoplastic GSH in resistant lines. The antioxidant cycle enzymes may function to remove apoplastic H₂O₂ with ascorbate and GSH derived from the cytoplasm. DHA and oxidized glutathione may be reduced in the apoplast or returned to the cytosol for rereduction.Abbreviations AA reduced ascorbate - APX ascorbate peroxidase - DHA dehydroascorbate (oxidised ascorbate) - DHAR dehydroascorbate reductase - G6PDH glucose-6-phosphate dehydrogenase - GSH reduced glutathione - GSSG glutathione disulphide - GR glutathione reductase - MDHA monodehydroascorbate - MDHAR monodehydroascorbate reductase - SOD superoxide dismutase     

Germination of Erysiphe gratninis f.sp. hordei conidia on barley leaves

Germination of Erysiphe graminis f.sp. hordei conidia on leaves of several barley cultivars was studied in the laboratory. On both detached leaves and intact plants, within 48 h of inoculation a higher proportion of conidia had germinated on the basal and middle portions of the adaxial leaf surface than on the corresponding portions of the abaxial surface. Such differences between surfaces were not observed near the leaf tip. Similar results were obtained with all the cultivars and growth stages tested, and with five isolates of E. graminis, and are consistent with the observation that there is usually less powdery mildew on the abaxial than the adaxial surface of barley leaves. With most of the barley genotype/mildew isolate combinations tested, within 48 h of inoculation higher proportions of conidia germinated on seedlings and juvenile plants than on older plants. Inherited characteristics which affect spore germination on the leaf surface may be important factors in the development of adult-plant resistance of barley to powdery mildew, particularly in certain genotypes.     

Apoplastic pH signaling in barley leaves attacked by the powdery mildew fungus Blumeria graminis f. sp. hordei

To investigate apoplastic responses of barley (Hordeum vulgare L.) to the barley powdery mildew fungus Blumeria graminis f. sp. hordei, noninvasive microprobe techniques were employed. H(+)- and Ca(2+)-selective microprobes were inserted into open stomata of barley leaves inoculated with Blumeria graminis f. sp. hordei race A6 conidia. Resistance gene-mediated responses of barley genotype Ingrid (susceptible parent line) and the near-isogenic resistant Ingrid backcross lines (I-mlo5, I-Mla12, and I-Mlg) were continuously monitored from 20 min to 4 days after inoculation. The main events were categorized as short-term responses around 2 h after inoculation (hai), intermediate responses around 8 and 12 hai, and long-term responses starting between 21 and 24 hai. Short-term responses were rapid transient decreases of apoplastic H(+)- and Ca2+ activities that lasted minutes only. Kinetics were similar for all genotypes tested, and thus, these short-term responses were attributed as nonspecific first encounters of fungal surface material with the host plasma membrane. This is supported by the observation that a microinjected chitin oligomer (GlcNAc)8 yielded similar apoplastic alkalinization. Intermediate responses are trains of H+ (increase) spikes that, being different in susceptible Ingrid and penetration-resistant I-mlo5 (or I-Mlg), were interpreted as accompanying specific events of papillae formation. Long-term events were massive slow and long-lasting alkalinizations up to two pH units above control. Since these latter changes were only observed with near-isogenic hypersensitive reaction (HR)-mounting genotypes I-Mla12 and I-Mlg but not with I-mlo5 or, to a smaller extent, with susceptible Ingrid, both lacking significant rates of HR, they were rated as cell death specific. It is concluded that apoplastic pH changes are important indicators of host-pathogen interactions that correlate with both the different stages of fungal development and the different types of host defense response.     

Expression in Escherichia coli and Simple Purification of Human Fhit Protein

The fragile histidine triad (Fhit) protein is a homodimeric protein with diadenosine 5′,5-P¹,P³-triphosphate (Ap₃A) asymmetrical hydrolase activity. We have cloned the human cDNA Fhit in the pPROEX-1 vector and expressed with high yield in Escherichia coli with the sequence Met-Gly-His₆-Asp-Tyr-Asp-Ile-Pro-Thr-Thr followed by a rTEV protease cleavage site, denoted as “H6TV,” fused to the N-terminus of Fhit. Expression of H6TV–Fhit in BL21(DE3) cells for 3 h at 37°C produced 30 mg of H6TV–Fhit from 1 L of cell culture (4 g of cells). The H6TV–Fhit protein was purified to homogeneity in a single step, with a yield of 80%, using nickel-nitrilotriacetate resin and imidazole buffer as eluting agent. Incubation of H6TV–Fhit with rTEV protease at 4°C for 24 h resulted in complete cleavage of the H6TV peptide. There were no unspecific cleavage products. The purified Fhit protein could be stored for 3 weeks at 4°C without loss of activity. The pure protein was stable at −20°C for at least 18 months when stored in buffer containing 25% glycerol. Purified Fhit was highly active, with a K_m value for Ap₃A of 0.9 μM and a k_cat(monomer) value of 7.2 ± 1.6 s⁻¹ (n = 5). The catalytic properties of unconjugated Fhit protein and the H6TV–Fhit fusion protein were essentially identical. This indicates that the 24-amino-acid peptide containing the six histidines fused to the N-terminus of Fhit does not interfere in forming the active homodimers or in the binding of Ap₃A.     

In vitro studies of the effect of Environmental Conditions on the anthacnose pathogen of bananas, Colletotrichum musae

In vitro studies were undertaken to determine the effect of pH, temperature, water availability and carbon dioxide (CO₂) concentration on germination and growth of Colletotrichum musae, the causal pathogen of anthracnose of bananas. The optimum pH for germination and growth varied between 4·0 and 5·0 depending on temperature. At low pH (< 3·0) and 15°C, both germination and growth were significantly reduced, with a marked increase in the lag time, in days, prior to growth. C. musae germinated and grew over a wide range of water activities (a_w; 0·995−0·94 and 0·995−0·92, respectively) at 20, 25 and 30°C. In all cases where germination occurred appresoria were subsequently produced. Optimum growth occurred at 30°C and 0·995 a_w, although this changed to 0·98 a_w at 35°C. Increasing CO₂ concentration to 15% or reducing oxygen concentration to 1% resulted in a significant (P < 0·05) reduction in growth, but did not inhibit growth completely.