APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jan. 1993, p. 1-60099-2240/93/010001-06$02.00/0Copyright © 1993, American Society for Microbiology

Identification and Ecology of Bacterial CommunitiesAssociated with Necroses of Three Cactus Species

JOAN L. M. FOSTER* AND JAMES C. FOGLEMAN2

Department ofBiology, Metropolitan State College ofDenver, Denver, Colorado 80217,1 andDepartment ofBiological Sciences, University ofDenver, Denver, Colorado 802082

Received 14 August 1992/Accepted 22 October 1992

To compare the bacterial communities residing in necrotic tissues of columnar cacti of the Sonoran Desert,isolates from 39 organ pipe, 19 saguaro, and 16 senita cacti were obtained. The isolates were clustered into 28conspecific groups on the basis of their fatty acid profiles. The distributions of the individual bacterial isolatesvaried among cactus species. Seven of the 28 species groups were unique to a particular cactus species, whereas8 species groups were found in all three cacti. The effective number of bacterial species for each cactus specieswas positively correlated with both the chemical complexity and glucose concentration of the plant tissues. Theeffective number of bacterial species and bacterial distribution patterns were compared with those known forcommunities of cactophilic yeasts. The observed bacterial distribution patterns are most likely due todifferences in the chemical compositions of the three cactus species.

The bacterial communities utilizing the necrotic tissues ofcolumnar cacti are important components of the cactus-microorganism-Drosophila model system of the SonoranDesert (2). Three of the columnar cactus species included inthe model system are organ pipe cactus (Stenocereus thurb-eni), saguaro cactus (Camegiea gigantea), and senita cactus(Lophocereus schottii). These three sympatric species ofcacti represent a favorable habitat for microbes in the midstof the arid desert environment. Injured cactus tissue can beinfected by bacteria from the phylloplane or blown in on dustand aerosols. This can result in the development of a cactusrot pocket or necrosis, which cactophilic yeasts can theninvade (14). Necrotic cactus tissues are used as substratesfor feeding and breeding by desert-adapted Drosophila spe-cies (15). The four Drosophila species that are endemic tothe Sonoran Desert are D. nigrospiracula, D. mojavensis, D.mettlen, and D. pachea. By flying from necrosis to necrosis,drosophilids serve as vectors for bacteria and yeasts (14, 18,19). Some of the desert-adapted drosophilids are restrictedto a specific cactus; i.e., D. pachea has unique nutritionalrequirements that limit it to feeding and breeding in necrosesof senita cactus. In contrast, D. mettleri can feed in all of thecolumnar cacti in its domain.These cactus species are chemically differentiated, espe-

cially with respect to secondary plant compounds. Forexample, organ pipe cacti contain high concentrations (ca.30%, dry weight) of triterpene glycosides (12). They alsocontain fatty acids with unusual chain lengths (16), predom-inantly C1o and C12, rather than the more typical plant fattyacid chain lengths of C16 and C18. In contrast, saguaro andsenita lack the triterpene glycosides and unusual-chain-length fatty acids but contain isoquinoline alkaloids that areabsent in organ pipe (22). Several of these plant compoundshave been shown to affect the microbes that reside innecrotic tissue (13, 30). Finally, fresh tissue of organ pipecacti has approximately three times the total carbohydratecontent of saguaro or senita tissue.The characterization of the types and frequencies of

bacteria in a community is a difficult task when standard

* Corresponding author.

techniques of bacterial identification are employed. Thisapproach to community ecology often requires that eachisolate be tested for its reaction to stains and for a multitudeof physiological characteristics. The data obtained fromthese tests can then be subjected to numerical taxonomy fordelineation of species clusters. Although this approach hasbeen used successfully by a number of investigators (20, 21,23), it does have the disadvantages of being expensive, laborintensive, and time consuming.Within the past decade, cellular fatty acid profiles have

been increasingly used to identify bacterial species (1, 5, 25,28). Since the fatty acid profile of a bacterial species,analyzed as fatty acid methyl esters (FAMEs), is unique tothat species, FAME analysis is an important identificationtool. This is evidenced by the fact that the major fatty acidsare included in many of the species descriptions in the ninthedition (1984) of Bergey's Manual of Systematic Bacteriol-ogy (24). However, this technique has not been widely usedto characterize bacterial communities from naturally occur-ring substrates.The established roles of bacteria in this well-studied model

ecological system include initiating the necroses. Bacteriaare the first microorganisms to grow in newly injured tissue,and cactophilic yeasts are secondary invaders (14). In addi-tion, the volatile products of bacterial fermentation of thecactus tissue are thought to be the chemical basis of hostplant selection by Drosophila species in the desert system(15). The mixtures of volatile compounds produced in necro-ses of different cactus species are distinct (12), and Dros-ophila species are attracted by these patterns in a species-specific manner. Bacteria (or the volatile compoundsproduced by them) may also serve as stimulants for ovipo-sition. Finally, both larval and adult stages of the life cycle ofthe cactophilic Drosophila species feed on bacteria. Thus,bacteria are important sources of nutrition.

Despite their central role in the cactus-microorganism-Drosophila model system, the bacterial component has notbeen as well characterized as the other components have. Inthis paper, we report the results of a survey of bacteriapresent in necrotic tissue of the three columnar cacti speciesmentioned above. Cellular fatty acid profiles are used tocluster isolates into conspecific groups. The bacterial com-

1

Vol. 59, No. 1

on October 27, 2020 by guest

http://aem.asm

.org/D

ownloaded from

2 FOSTER AND FOGLEMAN

munities among the different cacti are compared, and therelationships among the bacteria, the Drosophila species,and the chemistries of the cacti are discussed.

MATERIALS AND METHODS

Field work. Field work was conducted during March 1988in the Organ Pipe National Monument and in the vicinity ofTucson, Ariz. Necroses from 39 organ pipe, 19 saguaro, and16 senita cacti were sampled. All apparent stages of therotting process (e.g., young rots versus old) were included.Approximately 1 to 2 g of necrotic tissue was removed fromeach rot with a sterile knife and stored in a sterile 10-mlVacutainer tube on ice until it was sampled for bacteria.Tissue samples included surface material and extended intothe necrotic tissue to a depth of approximately 1 cm. Allsamples were streaked out within 10 h of collection.

Isolation of bacteria. Necrotic cactus tissue was streakedout on plates of Trypticase soy broth (TSB) agar containingcycloheximide (0.01%) and amphotericin B (0.001%) toinhibit fungal growth. The plates were incubated for 2 to 4days at ambient temperatures. All plates were examined,and colonies representing each distinct morphological typewere selected and streaked to single colonies three times onTSB agar plates.

Storage of bacterial isolates. Each pure isolate was grownovernight in 10 ml of TSB at 28°C. The bacteria weresedimented in a table-top centrifuge, the spent medium wasdecanted, and the pellet was suspended in TSB-glycerol (1:1)and stored at -15°C.

Conspecific grouping by whole-cell FAME profile. Thestored isolates were grown overnight at 28°C on TSB agar.They were then streaked on TSB agar (without the fungalinhibitors) and incubated at 28°C for 24 h. Colonies wereharvested from the quadrant containing confluent growthand spread on the bottom inside surface of a test tube (13 by100 mm). The bacterial lipids were saponified at 100°C for 30min in 15% NaOH dissolved in a 1:1 methanol-water solu-tion. The saponified samples were acidified with 60% HCI inmethanol, heated at 80°C for 10 minutes, and then cooledrapidly. The FAMEs were extracted in hexane-ether (1:1)for 10 min and then base washed for 5 min with 1.2% NaOH.The FAMEs in the organic phase were analyzed with aHewlett-Packard model 5890A gas chromatograph equippedwith a capillary column (cross-linked 5% phenyl methylsilicone; 25 m by 0.2 mm with film thickness of 0.33 ,um) anda flame ionization detector. The column conditions were asfollows: hydrogen carrier gas at 9 lb (ca. 4.1 kg) of headpressure and initial temperature of 170°C increasing at a rateof 5°C/min to a final temperature of 270°C. Data wereprocessed with a Hewlett-Packard 3392A integrator andanalyzed with a Hewlett-Packard 9133 personal computer.

Software originally purchased from Hewlett-Packard andsubsequently updated to version 3.0 by Microbial ID, Inc.,was employed to analyze the FAME profiles of the isolates.These analyses included cluster analysis by the unweightedpair-group method with arithmetic averages to constructconspecific groups of the isolates. Experience with multipleanalyses of many species indicates that isolates of the samespecies usually cluster together at a level of .10 Euclideandistance units (28). In this paper, a more conservative levelof 8 Euclidean distance units was used to define a species.Colony morphologies of isolates within a cluster were iden-tical or very similar. In ambiguous cases, the isolates werecompared by reactions in physiological tests as well. Oneisolate was selected as representative of each cluster. Selec-

tion of the representative was made on the basis of thegreatest similarity of its FAME profile with the largestnumber of isolates within the cluster. The identification ofthe representative of each cluster was based primarily onmorphology and physiological tests.

Identification by physiological characteristics. The pheno-typic characteristics of one representative isolate from eachconspecific group were determined by employing the testslisted in Table 1. All of the tests were performed by standardbacteriological procedures as described by Benson (3). Theresults were compared to the entries in the ninth edition ofBergey's Manual of Systematic Bacteriology, vol. 1 and 2(24).ENS. The effective number of species (ENS) of bacteria

was calculated for each cactus species by the method ofLaChance and Starmer (26). The equation for each cactusspecies is

n

ENS = 1/ z pi2i= 1

where p is the proportion of bacterial species i.

RESULTS

Conspecific grouping and identification of bacterial isolates.A dendrogram of the FAME profiles for all of the bacterialisolates from this survey was constructed by cluster analy-sis. When essentially the same FAME profile was observedfor two or more isolates from the same sample, then all butone of those isolates were eliminated as redundant. The 27isolates that were detected in two or fewer samples wereexcluded because of their low frequency. The remainingisolates formed 28 conspecific groups on the basis of simi-larity of FAME profiles. One member of each conspecificgroup was then identified on the basis of traditional taxo-nomic methods. The results of the taxonomic tests aresummarized in Table 1. All but 2 of the 28 representativeisolates were identified to at least the genus level, and 15were identified to the species level. The putative identitygiven in the table represents the closest match to speciesdescribed in Bergey's Manual of Systematic Bacteriology(24). As expected, gram-negative species outnumber gram-positive species, and a large portion of the isolates aremembers of the family Enterobacteriaceae. Figure 1 is thedendrogram constructed from fatty acid data and shows therelationships among the 28 isolates identified in Table 1.

Bacterial distribution patterns. Of the 28 conspecificgroups, 20 were detected in the organ pipe necroses, 19 weredetected in the saguaro necroses, and 18 were detected in thesenita necroses (Table 2). The calculated ENS values fororgan pipe, senita, and saguaro are 15.2, 12.6, and 10.3,respectively. The predominant bacterial species was Xanth-omonas fragariae, which was isolated with high frequencyfrom all cactus species; Pseudomonas was the genus repre-sented by the most species (seven).

Eight bacterial species (29%) were isolated from all threecacti, whereas seven species (25%) were only found innecroses of one cactus. The remaining 13 species (46%) wereisolated from two of the three cacti. The distribution of these13 species appears random with respect to the pairs of cactusspecies (i.e., 5 species in organ pipe and saguaro; 5 speciesin saguaro and senita; 3 species in organ pipe and senita).

APPL. ENvIRON. MICROBIOL.

on October 27, 2020 by guest

http://aem.asm

.org/D

ownloaded from

BACTERIAL COMMUNITIES IN NECROTIC CACTUS TISSUE 3

-4 0.\ LA 4.6 \0 --I LAt -P- W 0

0.0.0. 0 0 0.00 0 0.0. a00.a000 .0

0 ,) q U)

lil ++ l+++ ++

+ + + t+ + + + + + + + + + + III+

+++I++++++I+I I+++++I+++

+++ I ++++++I I+++++ I I

Z+0I+++

I+ +II

+ I + ++ + ++

++++++++++++I I++++++++++I+++

I+ II++I I++++++I I+ II

+

I +

+

+ + + +

+++I+I I+I++I+++I

I +I + + I I

++

Z

0

0

0

+++ ++ + I+++ + +

++

++

I I IZ I

IZ I

I I IZ I

+

+ +

++ +

+

+ + +

z40

z0:

z

z

z0

+++ I IIII II 1+1 II+1I I I lIIUI

00~~~~~~~ R Ro'o6 o on

o oQn.C kti

rA o D-A9r ntkkk

CD D ) C.

c c). 2.O. o Co CO COCaC C (A

~ ~ ~~D ;

0

0

0.

a,

CA

Pz

u)=D

Z

CD

a,

aI0

o D

CD

a,

gD0

CDCD

CD

a,0

c0a,

CD

VDa,

+ +++

I1I+

VOL. 59, 1993

CDD q oN

C

CD2

a, a, X.5-

n 8 M 0 0

C o 0a

o CO

_0. cn

CDa

Q:rA

oa,

o 0.C

a,.

CD

In

oD

aCL

ao

0

0.

5.a,

CD

5.

CD

CD0.

CD

0.

On

++

+ z

t.110a,.

5a,coIt

0

IDN.

a,0

a,o

CD

CDItMIO57CDcn

+ + + + +

on October 27, 2020 by guest

http://aem.asm

.org/D

ownloaded from

4 FOSTER AND FOGLEMAN

ClusterNumber

1 -

2-

3-

4-

5-

6-

Euclidian Distance8.1 16.2 24.3 32.4 40.5 48.6 56.8 64.9

I-

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28I I

FIG. 1. Dendrogram of the 28 conspecific groups isolated fromnecrotic cactus tissue.

DISCUSSION

In this study, the fatty acid composition of bacteria wasused to construct clusters of conspecific isolates, which werethen identified by traditional taxonomic methods. Althoughthe FAME patterns are specific to each bacterial species anda commercial library for identifying bacteria by their FAMEpattern exists, this library is not always applicable to envi-ronmental isolates for several reasons. First, the commerciallibraxy is constructed from FAME analysis of type and stockcultures that are primarily of clinical interest. The commer-cial library contains few environmental isolates and does notcontain any isolates from cactus necroses. Additionally, oneof the bacteria most frequently isolated from cactus rots,Erwinia cacticida, was only recently described (1) and is notyet in the commercial library. Most likely, several other

TABLE 2. Frequency of isolation of bacterial species fromcactus necroses

Frequency of isolation (%) from:Cluster Puaieseisnmno. Putative species name Organ pipe Saguaro Senita

(n = 39) (n = 19) (n = 16)

1 Pseudomonas species 5 0 11 62 Pseudomonas species 2 10 5 63 Staphylococcus aureus 0 16 194 Enterococus faecium 10 11 05 Pseudomonas species 1 5 5 06 Xanthomonas ampelina 10 0 137 Pseudomonas species 6 13 0 08 Pseudomonas species 4 13 0 09 Stomatococcus species 5 0 610 Staphylococcus species 2 26 0 011 Staphylococcus species 1 18 16 012 Serratia marcescens 5 5 613 Citrobacterffreundii 10 11 1914 Erwinia species 1 8 5 615 Unknown species 1 5 11 016 Unknown species 2 5 5 017 Xanthomonas fragariae 23 53 3818 Pseudomonas putida 0 6 1319 Erwinia mallotivora 8 0 020 Erwinia uredovora 10 5 621 Erwinia species 2 0 0 2522 Erwinia chrysanthemi 0 11 623 Erwinia stewartii 0 0 2724 Edwardsiella hoshinae 18 5 625 Pseudomonas species 3 0 11 026 Erwinia cacticida 23 32 3127 Erwinia ananas 0 5 1928 Erwinia herbicola 26 0 25

isolates included in our collection represent undescribedspecies. The situation with the cactophilic yeast communityis comparable. Starmer et al. (33) reported that 12 of the 33commonly isolated cactophilic yeasts are not found outsideof the cactus-Drosophila niche.For microbial ecology studies, the strength of the FAME

method is the ability to identify conspecific groups of isolatesby clustering analysis. Actual species designations are notnecessary to make comparisons as to where and when eachbacterial species is detected in different substrates (e.g.,cactus species). FAME analyses enable communities to becompared without the time-consuming task of trying toidentify each bacterial isolate to the species or even thegenus level. Grouping isolates by physiological profile wouldbe prohibitively time consuming as well.

Other investigators have used fatty acid analysis forcharacterization of whole microbial communities in naturalsettings. This is possible because many classes of organismscontain unique fatty acids (11). Measurements of the fattyacids pooled from the whole community provide a basis forestimations of percent bacteria, cyanobacteria, diatoms, andfungi (5, 10). Fatty acid analyses of communities have beenused to document eukaryotic-prokaryotic relationships inbenthic communities (5) and shifts in the community com-positions of estuarine detrital microbiota (4).

Fatty acid analyses have also been used to identify indi-vidual bacterial species and even subspecies (6). The cultureage, pH of growth medium, and carbohydrate source allaffect the fatty acid composition of bacteria (9). However,with standardized culture conditions, fatty acid profiles are arapid, reproducible method of classifying bacteria (7-9, 27,

APPL. ENVIRON. MICROBIOL.

on October 27, 2020 by guest

http://aem.asm

.org/D

ownloaded from

BACTERIAL COMMUNITIES IN NECROTIC CACTUS TISSUE 5

34). As such, the major fatty acids are routinely used in theidentification of clinical isolates as well as in new descrip-tions of bacterial species (1, 24).

In this survey of 74 cactus necroses, cluster analysis of thefatty acid profiles of bacterial isolates resulted in the con-struction of 28 conspecific groups of bacteria. A comparablenumber of yeast species (i.e., 26 species) were previouslyisolated in surveys of cactus necroses (33). Although the useof 8 Euclidian distance units (Fig. 1) to separate species issomewhat arbitrary, the fact that isolates of the conspecificgroups were distinguishable on the basis of other character-istics (e.g., physiological tests) is supportive. It is recognizedthat the number of species recovered from these substratesis limited to bacteria that are culturable under the conditionsemployed. Although bacterial densities in cactus rots deter-mined under aerobic versus anaerobic conditions were notsignificantly different (17b), species composition could bedifferent. In addition, fastidious organisms that cannot growon TSB agar would not have been isolated.The number of isolates in any particular conspecific group

varied, ranging up to 37 in the cluster representing X.fraganae. Several factors may contribute to the variation innumbers of isolates per bacterial species. One possibility isthat some bacteria are better adapted to the cactus necrosisniche and are more tolerant of potentially toxic secondaryplant compounds (e.g., triterpene glycosides and isoquino-line alkaloids). It is also possible that the less frequentlyisolated species represent bacteria from different succes-sional stages. A species occurring in an early successionalstage in necrosis development may be prone to microbialantagonism or competition and eliminated from later stages.Thus, bacteria from early successional stages would only bepresent briefly and would be less likely to be sampledbecause young rots are less conspicuous. This last explana-tion corresponds with observations on yeast communitystructure in both organ pipe and agria (Stenocereus gummo-sus) cactus necroses. Starmer (29) and Fogleman andStarmer (17) examined yeast distributions among differentnecroses and concluded that variations in yeast speciesreflected different successional states.The ENS is a more informative ecological parameter than

just counting the number of species isolated, because ENSreflects the proportion of the community each species rep-resents. Thus, ENS takes into account unequal sample sizesand the evenness with which various bacteria are repre-sented in the samples within each cactus species. Kircher(22) ranked the general chemical complexity of these threecacti as organ pipe > senita > saguaro. Since the rank orderof ENS is identical to the chemical complexity, there is anapparent positive correlation between these two variables;i.e., the more chemically complex tissue supports a greaterENS. This correlation, however, does not extend to thecactophilic yeasts. The ENS values calculated from the dataon frequency of yeast isolation reported by Starmer andFogleman (31) do not correlate with the chemical complex-ity, since the ENS ranking (ENS value) for yeast species issaguaro (8.7) > organ pipe (8.2) > senita (3.0).The data in Table 2 demonstrate that the bacterial com-

munities of the three species of cactus are different. Onlyeight of the 28 isolated bacterial species were detected in allthree cacti. The remaining 20 bacterial species were detectedin only one (7 of 28) or two (13 of 28) of the cactus types. Itappears that the eight bacterial species isolated from all threetypes of cactus are not fastidious but can utilize the differentsubstrates and growth conditions present. Of those eightspecies, X. fraganiae and E. cacticida were the most fre-

quently isolated bacteria, suggesting that they are general-ists.A similar pattern of cactus utilization has been observed

with the yeasts surveyed in the same cacti by Starmer andFogleman (31). In their survey, 6 of the 24 yeasts wereisolated from all three types of cactus necroses. In a reviewpaper, Starmer et al. (33) concluded that differences in hostplant chemistry contribute significantly to yeast distributionpatterns. For example, Pichia amethionina var. pachycere-ana was not detected in organ pipe cacti, and this species hasbeen shown to be inhibited by organ pipe triterpene glyco-sides (31). They also hypothesized that vectors such ascactophilic Drosophila species may contribute to the yeastdistribution patterns.A number of different factors could contribute to the

observed bacterial distribution patterns. One factor could bethe availability of specific nutrients in the plant tissues. Thisconclusion would concur with the positive correlation be-tween ENS and chemical complexity. The greater the vari-ety of compounds (potential nutrients), the greater thevariety of bacterial species. In addition to the variety ofnutrients, the availability of carbohydrates could influencethe number of bacteria capable of utilizing cactus tissue as asubstrate. Fogleman and Abril (12) determined the glucoseconcentrations in fresh cactus tissue to be 3.36, 1.90, and1.24% (dry weight) for organ pipe, senita, and saguaro,respectively. This, too, shows a positive correlation withbacterial ENS.

Distribution patterns could also be partly explained ifcertain bacteria were dependent on Drosophila species fortransport from plant to plant, since the flies themselves arecactus specific. However, Fogleman and Foster (14) andFoster (17a) demonstrated that bacterial colonization ofnewly injured agria cactus tissue was not hindered by theexclusion of Drosophila species and other large insect vec-tors.Another determining factor for the observed bacterial

distribution patterns involves inhibition by plant allelochem-icals, which differ among cactus species (22). Other re-searchers have determined that cactus allelochemicals doaffect cactophilic yeasts. For example, Pichia amethioninavar. pachycereana and Pichia opuntiae var. opuntiae havebeen shown to be inhibited by powdered tissue from organpipe but not by powdered tissue from senita or saguaro (32).Additionally, Starmer (30) demonstrated that the growth offive cactophilic yeasts was inhibited by adding sodium saltsof two fatty acids (i.e., caproic and caprylic) derived fromorgan pipe cactus. The effects of such allelochemicals onbacterial species, however, have yet to be tested.

In conclusion, although the organ pipe, senita, andsaguaro cacti are sympatric, their necroses do not have thesame microbial communities. Because the bacteria havebeen shown to be independent of drosophilids acting asvectors, the observed bacterial patterns cannot be explainedby fly utilization patterns. The distribution patterns mayreflect the chemical complexity of the substrates, availablecarbohydrates, or selective growth inhibition of bacterialspecies. Plant allelochemicals restrict the growth of somecactophilic yeast species (29, 31) and may restrict the growthof some species of bacteria.

ACKNOWLEDGMENTS

We gratefully acknowledge the technical assistance of Gregory A.Price and Xuan-Trang Truong. We also thank Harold J. Smith,Superintendent of the Organ Pipe National Monument (NationalPark Service), for the use of park facilities and permission to collect.

VOL. 59, 1993

on October 27, 2020 by guest

http://aem.asm

.org/D

ownloaded from

6 FOSTER AND FOGLEMAN

This research was supported by Public Health Service grant GM34820 from the National Institute of General Medical Sciences toJ.C.F.

REFERENCES1. Alcorn, S. M., T. V. Orum, A. G. Steigerwalt, J. L. M. Foster,

J. C. Fogleman, and D. J. Brenner. 1991. Taxonomy andpathogenicity of Erwinia cacticida sp. nov. Int. J. Syst. Bacte-riol. 41:197-212.

2. Barker, J. S. F., and W. T. Starmer. 1982. Ecological geneticsand evolution: the cactus-yeast-Drosophila model system. Ac-ademic Press, Australia.

3. Benson, H. J. 1990. Microbiological applications. William C.Brown Co., Publishers, Dubuque, Iowa.

4. Bobbie, R. J., J. S. Nickels, G. A. Smith, S. D. Fazio, R. H.Findlay, W. M. Davis, and D. C. White. 1981. Effect of light onbiomass and community structure of estuarine detrital micro-biota. Appl. Environ. Microbiol. 42:150-158.

5. Bobbie, R. J., and D. C. White. 1980. Characterization ofbenthic microbial community structure by high resolution gaschromatography of fatty acid methyl esters. Appl. Environ.Microbiol. 39:1212-1222.

6. DeBoer, S. H., and M. Sasser. 1986. Differentiation of Erwiniacarotovora ssp. carotovora and E. carotovora ssp. atrosepticaon the basis of cellular fatty acid composition. Can. J. Micro-biol. 32:796-800.

7. Dees, S. B., and C. W. Moss. 1975. Cellular fatty acids ofAlcaligenes and Pseudomonas species isolated from clinicalspecimens. J. Clin. Microbiol. 1:414-419.

8. Dees, S. B., and C. W. Moss. 1978. Identification of Achromo-bacter species by cellular fatty acids and by production of ketoacids. J. Clin. Microbiol. 8:61-66.

9. Drucker, D. B. 1976. Gas-liquid chromatographic chemotaxon-omy, p. 52-125. In J. R. Norris (ed.), Methods in microbiology.Academic Press, London.

10. Federle, T. W., M. A. Hullar, R. J. Livingston, D. A. Meeter,and D. C. White. 1983. Spatial distribution of biochemicalparameters indicating biomass and community composition ofmicrobial assemblies in estuarine mud flat sediments. Appl.Environ. Microbiol. 45:58-63.

11. Federle, T. W., and D. C. White. 1982. Preservation of estuarinesediments for lipid analysis of biomass and community structureof microbiota. Appl. Environ. Microbiol. 44:1166-1169.

12. Fogleman, J. C., and J. R. Abril. 1990. Ecological and evolu-tionary importance of host plant chemistry, p. 121-143. InJ. S. F. Barker, R. J. Maclntyre, and W. T. Starmer (ed.),Ecological and evolutionary genetics of Drosophila. PlenumPublishing Corp., New York.

13. Fogleman, J. C., and L. A. Armstrong. 1989. Ecological aspectsof cactus triterpene glycosides. I. Their effect on fitness com-ponents of Drosophila mojavensis. J. Chem. Ecol. 15:663-676.

14. Fogleman, J. C., and J. L. M. Foster. 1989. Microbial coloniza-tion of injured cactus tissue (Stenocereus gummosus) and itsrelationship to the ecology of cactophilic Drosophila mojaven-sis. Appl. Environ. Microbiol. 55:100-105.

15. Fogleman, J. C., and W. B. Heed. 1989. Columnar cacti anddesert Drosophila: the chemistry of host plant specificity, p.1-24. In J. Schmidt (ed.), Special biotic relationships in the aridsouthwest. University of New Mexico Press, Albuquerque.

16. Fogleman, J. C., and H. W. Kircher. 1986. Differential effects offatty acid chain length on the viability of two species ofcactophilic Drosophila. Lipids 21:92-99.

17. Fogleman, J. C., and W. T. Starmer. 1985. Analysis of thecommunity structure of yeasts associated with the decaying

stems of cactus. III. Stenocereus thurbeni. Microb. Ecol. 11:165-173.

17a.Foster, J. L. M. 1990. Ph.D. thesis. University of Denver,Denver, Colo.

17b.Foster, J. L. M., and J. C. Fogleman. Unpublished data.18. Ganter, P. F. 1988. The vectoring of cactophilic yeasts by

Drosophila. Oecologia 75:400-404.19. Ganter, P. F., W. T. Starmer, M.-A. Lachance, and H. J. Phaff.

1986. Yeast communities from host plants and associated Dros-ophila in southern Arizona: new isolations and analysis of therelative importance of hosts and vectors on community compo-sition. Oecologia 70:386-392.

20. Holder-Franklin, M. A., M. Franklin, P. Cashion, C. Cornier,and L. Wuest. 1978. Population shifts in heterotrophic bacteriain a tributary of the Saint John's river as measured by taxomet-rics, p. 45-50. In M. W. Loutit and J. A. R. Miles (ed.),Microbial ecology. Springer Verlag, Berlin.

21. Kaneko, T., R. M. Atlas, and M. Krichevsky. 1977. Diversity ofbacterial populations in the Beaufort Sea. Nature (London)270:596-599.

22. Kircher, H. W. 1982. Chemical composition of cacti and itsrelationship to Sonoran Desert Drosophila, p. 143-158. InJ. S. F. Barker and W. T. Starmer (ed.), Ecological genetics andevolution: the cactus-yeast-Drosophila model system. Aca-demic Press, Australia.

23. Kolbel-Boelke, J., B. Tienken, and A. Nehrkorn. 1988. Microbialcommunities in the saturated groundwater environment. I.Methods of isolation and characterization of heterotrophic bac-teria. Microb. Ecol. 16:17-29.

24. Krieg, N. R., and J. G. Holt. 1984. Bergey's manual of system-atic bacteriology, vol. 1 through 4. The Williams and WilkinsCo., Baltimore.

25. Kuykendall, L. D., M. A. Roy, J. J. O'Neill, and T. E. DeVine.1988. Fatty acids, antibiotic resistance, and deoxyribonucleicacid homology groups of Bradyrhizobium japonicum. Int. J.Syst. Bacteriol. 38:358-361.

26. LaChance, M.-A., and W. T. Starmer. 1982. Evolutionarysignificance of physiological relationships among yeast commu-nities associated with trees. Can. J. Bot. 60:285-293.

27. Lechevalier, M. P. 1977. Lipids in bacterial taxonomy-a tax-onomist's view. Crit. Rev. Microbiol. 7:109-209.

28. Sasser, M. 1990. Tracking a strain using the microbial identifi-cation system. MIDI technical note 102. MIDI, Newark, Del.

29. Starmer, W. T. 1982. Analysis of community structure of yeastsassociated with the decaying stems of cactus. I. Stenocereusgummosus. Microb. Ecol. 8:71-81.

30. Starmer, W. T. 1982. Associations and interactions amongyeasts, Drosophila and their habitats, p. 159-174. In J. S. F.Barker and W. T. Starmer (ed.), Ecological genetics and evo-lution: the cactus-yeast-Drosophila model system. AcademicPress, Sydney, Australia.

31. Starmer, W. T., and J. C. Fogleman. 1986. Coadaptation ofDrosophila and yeasts in their natural habitat. J. Chem. Ecol.12:1037-1055.

32. Starmer, W. T., H. W. Kircher, and H. J. Phaff. 1980. Evolutionand speciation of host plant specific yeasts. Evolution 34:137-146.

33. Starmer, W. T., H. J. Phaff, M. Miranda, M. W. Miller, andW. B. Heed. 1982. The yeast flora associated with the decayingstems of columnar cacti and Drosophila in North America.Evol. Biol. 14:269-294.

34. Suzuki, K.-I., and K. Komagata. 1983. Taxonomic significanceof cellular fatty acid composition in some coryneform bacteria.Int. J. Syst. Bacteriol. 33:188-200.

APPL. ENVIRON. MICROBIOL.

on October 27, 2020 by guest

http://aem.asm

.org/D

ownloaded from