Vol. 58, No. 12APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 1992, p. 3914-39200099-2240/92/123914-07$02.00/0Copyright © 1992, American Society for Microbiology

Heterotrophic Bacteria in an Air-Handling SystemPHILIP HUGENHOLTZ AND JOHN A. FUERST*

Centre for Bacterial Diversity and Identification, Department ofMicrobiology,University of Queensland, Brisbane, Queensland 4072, Australia

Received 3 August 1992/Accepted 29 September 1992

Heterotrophic bacteria from structural surfaces, drain pan water, and the airstream of a well-maintainedair-handling system with no reported building-related illness were enumerated. Visually the system appearedclean, but large populations of bacteria were found on the fin surface of the shpply-side cooling coils (10 to 106CFU cm-2), in drain pan water (10' to 107 CFU ml-'), and in the sump water of the evaporative condenser(10S CFU ml-). Representative bacterial colony types recovered from heterotrophic plate count cultures onR2A medium were identified to the genus level. Budding bacteria belonging to the genus Blastobacterdominated the supply surface of the coil fins, the drain pan water, and the postcoil air. These data andindependent scanning electron microscopy indicated that a resident population of predominantly Blastobacterbacteria was present as a biofilm on the supply-side cooling coil fins.

The microbiology of building air-handling systems (AHSs)first came to attention in the 1970s because of health prob-lems, such as humidifier fever and legionellosis, arising frommicrobial contamination of components of the AHSs, par-ticularly humidification systems and cooling water (3, 8, 30).The problem was exacerbated in the 1980s by the tighteningup of buildings to conserve energy, which coincided withincreased reports of a range of building-related health prob-lems collectively known as sick-building syndrome (12).Sick-building syndrome has been determined to be mostdirectly related to inadequate ventilation and the consequentbuildup of airborne pollutants, including microorganismsand nonbiological factors such as radon, formaldehyde, andcarbon monoxide (19, 26). In modern buildings, microbio-logical sources may account for only a minor proportion ofbuilding-related health problems, in the order of 5% accord-ing to one estimate (26). However, this small proportionencompasses the most debilitating forms: allergic or hyper-sensitivity reactions and infections (including legionellosis)collectively termed building-related illness (26). Apart frombuilding-related illness, biofouling can also cause engineer-ing problems in AHSs. Heat-exchange systems are known tobe adversely affected by microbial biofilms in terms ofreduced heat transfer rates and premature replacement ofequipment because of microbiologically induced corrosion(4-6).

Little is known about the types of microorganisms com-monly encountered in AHSs, and only preliminary cutoffthresholds for levels of microorganisms considered to behazardous to public health have been proposed (20, 26).Organizations such as the National Institute for Occupa-tional Safety and Health have been the chief sources of dataconcerning AHS microflora (16, 20), but such data haveoften been obtained on request from buildings demonstratingbuilding-related illness and usually focus on specific problemareas such as contaminated cooling waters or humidificationsystems, rather than the entire system.The present study documents heterotrophic bacterial

numbers in an AHS with no record of building-related illnessto provide data on the baseline bacterial levels present in a"healthy" AHS. Identification of bacteria to genus level in

* Corresponding author.

certain samples was also performed to increase knowledgeof the types of bacteria commonly encountered in the AHS.

MATERIALS AND METHODS

The AHS. The AHS under study consisted of two rela-tively large built-up air-handling units (AHUs), housed inseparate roof plant rooms, servicing a five-story library.Each AHU contained extended-surface disposable filters, adouble-inlet backward curve centrifugal fan, electric reheatbanks, and a direct expansion evaporator (cooling) coil (25cm deep) with split-level drain pans forming continuouscircuits around the base of the coil banks. The cooling coilswere serviced by three evaporative condensers and threereciprocating open-drive compressors. The system wasmaintained in accordance with Australian standard guide-lines (28). Sump water from the evaporative condenser wasdose treated with a commercial alkaline biocide (BiocideP109; Maxwell Chemicals Ltd., Botany, New South Wales,Australia) and treated on line (conductivity feedback con-trol) with an anticorrosive agent (Corrostop CB; MaxwellChemicals Ltd., Botany, New South Wales, Australia).Because of the subtropical climate of Brisbane, the AHUsdid not contain humidifier systems and the ducting was notinsulated.

Sampling of the AHS. The AHS (predominantly AHU 1)was sampled on several occasions from September 1989through September 1990. A schematic diagram of AHU 1and the closest evaporative condenser (located 2 m from thefresh-air intake ofAHU 1) is shown in Fig. 1. Samples wereobtained from the sites indicated in Fig. 1. On some samplingoccasions, multiple samples were obtained from differentareas of the same site.Air samples were taken with a Sartorius MD-8 air sampler

fitted with sterile gelatin filters. A volume of air (1 mi3) wasdrawn through the filter (8 min at a flow rate of 8 m h-1), andthen the filter was dissolved in 9 ml of prewarmed sterile 0.01M phosphate-buffered saline (PBS). Water samples (50 to100 ml) were collected in sterile 120-ml Whirlpak bags.Surface samples (9 cm2) were taken with sterile cotton swabsand transferred back to the laboratory in 9 ml of sterile 0.01M PBS. After the initial sampling occasion in September1989, surface samples of the cooling coils were obtained with

3914

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from

brought to you by COREView metadata, citation and similar papers at core.ac.uk

provided by University of Queensland eSpace

HETEROTROPHIC BACTERIA IN AN AIR-HANDLING SYSTEM 3915

FIG. 1. Schematic drawing ofAHU 1 and the associated evapo-

rative condenser. The direction of airflow is indicated by arrows.

Sampling sites are indicated by abbreviations. Air samples: Al,return-fresh air mix; A2, precoil air; A3, postcoil air; A4, supply airregister most remote from AHU 1. Water samples: Wl, evaporativecondenser sump water; W2, drain pan water (continuous circuitaround base of coils). Surface samples: S1, air intake grill; S2,filters, return-side surface (return filters); S3, filters, supply-sidesurface (supply filters); S4, fan chamber surface; S5, cooling coilfins, return side (return coils); S6, cooling coil fins, supply side(supply coils); S7, duct surface 2 meters past coils (postcoil duct);S8, supply air register most remote from AHU 1.

Cytobrushes (Medscand, Sweden) because of their durabil-ity and closer fit to the fin gaps.

Heterotrophic plate count. All samples were serially di-luted in sterile 0.01 M PBS. Bacteria were enumerated induplicate by the spread plate technique with R2A agar (22)and incubated at 22°C for 1 week. For air samples, 15-cm-diameter plates of R2A agar were used, since counts wereexpected to be low. Individual colony morphology typeswere enumerated from plates with 20 to 300 colonies. Watersamples were tested for the presence of Legionella spp. onBCYE agar (10) and BMPA agar (9) on the first samplingoccasion in September 1989.

Isolation and identification of representative colony mor-phology types. For some samples, representative colonymorphology types were subcultured onto R3A agar (22) fromthe heterotrophic plate count R2A plates (7 days old) con-taining 20 to 300 colonies and incubated at 22°C. Multipleselections of colony types, depending on the total number ofcolonies for a given morphology type, were taken: fewerthan 20 colonies, one isolate; 20 to 50 colonies, two isolates;50 to 100 colonies, three isolates; more than 100 colonies,four isolates. The numbers of colonies for each colonymorphology type were counted, and representatives of eachcolony type were identified and assigned to a bacterialgenus. From these determinations a count was derived foreach bacterial group. This protocol incorporated the as-

sumption that all colonies of a particular type belonged to thesame bacterial genus.

Isolates were initially characterized to the genus level asdescribed by LeChevallier et al. (18) with the followingcharacteristics: Gram stain, colony morphology on R3A me-

dium, pigmentation, cell morphology, motility in R3A broth,

growth on R3A at 35°C, oxidation-fermentation of glucose ina medium designed for nonfermentative gram-negative rods(21), and cytochrome oxidase and indole production. Refer-ence strains for these tests, obtained from the AustralianCollection of Microorganisms, University of Queensland,were Moraxella lacunata (ACM647, ATCC 17952), Acineto-bacter calcoaceticus (ACM615, ATCC 17912), Pseudomonascepacia (ACM1771, ATCC 25416), Alcaligenes faecalis(ACM503, ATCC 19018), Flavobactenum meningosepticum(ACM3094, ATCC 13253), Flavobacterium odoratum (ACM3097, ATCC 29979), Aeromonas hydrophila (ACM2433,ATCC 7966), Escherichia coli (ACM1803, ATCC 11775),Staphylococcus aureus (ACM556, ATCC 9144), Micrococcusluteus (ACM975), Bacillus subtilis (ACM40, ATCC 6051),Arthrobacter globiformis (ACM2455, ATCC 8010), and Co-rynebacterium xerosis (ACM2476, ATCC 373). Isolates capa-ble of producing visible colonies on R3A within 2 days at 35°Cwere further characterized by using a Vitek Senior system(Vitek Systems, Hazelwood, Mo.). Isolates were stored at-60°C in R3A medium supplemented with 10% glycerol.Absorption spectra of bacterial pigments were obtained

from methanol extracts of selected strains of R3A agar-grown cells with a Hitachi 150-20 double-beam spectropho-tometer.

Transmission electron microscopy. Cells from 3-day-oldplate cultures were negatively stained with 1% uranyl ace-tate in 0.4% sucrose. Grids were examined on a HitachiH-800 transmission electron microscope at 100 kV.For thin sections, plate cultures of selected isolates were

fixed with glutaraldehyde and OS04 in cacodylate buffer,dehydrated in a graded ethanol series, and embedded in LRwhite resin. Thin sections stained with lead citrate anduranyl acetate were examined with a Hitachi H-800 trans-mission electron microscope.

Scanning electron microscopy. Small sections (ca. 2 cm2) ofcoil fin were removed directly from the cooling coil bankwith flame-sterilized tin snips. Stubs (6 mm in diameter)were punched from the sections with a flat-ended bit, fixed in2.5% glutaraldehyde in 0.1 M cacodylate buffer for 30 min,and then washed in the same buffer (three times for 5 mineach). The stubs were dehydrated in a graded ethanol series,critical-point dried, and sputter coated with gold. Stubs wereobserved on a Phillips 505 scanning electron microscope.

RESULTS

Total heterotrophic counts. Heterotrophic bacteria wereenumerated from 13 sites within the test AHS and one watersample from an associated evaporative condenser in Sep-tember 1989 with follow-up sampling through September1990 (Table 1).The highest numbers of bacteria were found in the water

samples and on the supply-side surface of the cooling coils.As estimated from colonies developing on R2A, fungi tendedto dominate on dry surfaces such as the fan chamberhousing, duct surface, and some parts of the return-sidecooling coils; no bacteria were detected at these sites. Thereturn-side cooling coil surface had highly variable microbialpopulations, both spatially and temporally, ranging from nodetectable bacteria to 105 bacterial CFU cm-2 (Table 1).Conversely, the supply coil surface had consistently highbacterial numbers (mostly between 105 and 106 CFU cm-2)over the sampling period and at different sampling areas onthe supply coil surfaces of both AHUs.

Air samples contained low levels of bacteria, with theexception of the postcoil air sample taken in September

VOL. 58, 1992

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from

3916 HUGENHOLTZ AND FUERST

TABLE 1. Summary of bacterial counts taken from sites in thetest AHS between September 1989 and September 1990

NoofbctraSampling sitea Date No. of bacterial

CF

AirReturn-fresh air mix

(Al)Precoil air (A2)Postcoil air (A3)

Supply air (A4)

WaterEvaporativecondenser sumpwater (Wi)

Drain pan water (W2)

SurfacesAir intake grill (S1)

Return filters (S2)Supply filters (S3)Fan chamber (S4)Return coils (S5)

Return coils (AHU 2)

Supply coils (S6)

Supply coils (AHU 2)

Postcoil duct (S7)Supply air register

(S8)

September 1989

April 1990September 1989April 1990b

September 1989

September 1989

January 1990September 19891 February 19905 February 1990April 1990"

September 1989January 1990February 1990February 1990February 1990September 1989March 1990"

June 1990September 1990June 1990September 1990January 1990February 19906 April 1990b

12 April 1990b

June 1990b

September 1990June 1990b

Septemberlggob

February 1990September 1989

February 1990

m-3'

105 Mm3

1,640 m-3c155 Mm3

140 Mm3m-3'

1.6 x 10 ml-lc

105ml-'105ml-lc

4.9 x 107 mI-16.3 x 107 mI-11.0 x106 mI-1

105ml1

1,050 cm-2cNDd

190 cm-2280 cm-2

NDcm-2n

2.3 x 10 cm-2NDND

7.5 x104cm-21,600 cm-2<1.0 x 04 cm-2

ND105cm-2-

>3,000 cm-21.1 x 07 cm-26.4 iO cm-2c9.0 X 104 cm-26.0 x 104 cm-28.0 x 105cm-2c

105cm-2A6.6 x 104 cm-21.0 x 105 cm-2A3.3 x 105 cm-21.7 x 106 cm-2c1.7 x 106 cm-2c6.5 x 105 cm-2

4.0 x 105cm-2NDcm-2c

100 cm-2

a Abbreviations within parentheses refer to sites shown in Fig. 1 or tocomparable sites in AHU 2.

b Multiple samples were obtained from different areas of the same site onthis date.

I Representative colony types were subcultured and identified.d ND, not detected.

1989, which had a bacterial count of 103 CFU m-3. This wasin the order of 10 times greater than the postcoil air bacterialcount obtained in April 1990 (Table 1) and was most proba-bly due to the presence of condensate on the coils in

September but not in April. Bacteria in the condensateaerosol across the air sampler would account for the ele-vated bacterial counts. The filters sampled had low counts ofbacteria relative to those on other surfaces and even relativeto air samples. This was consistent with the filters beingreplaced shortly before the time of sampling.

Bacterial profiles. Bacteria were identified to the genuslevel for a selection of sampling sites in the test AHSpredominantly on the first sampling occasion in September1989 (Table 2).A pink-pigmented budding bacterium belonging to the

genus Blastobacter (see below for a more detailed descrip-tion) was recovered in high numbers from the supply coils(105 to 106 cm-2) and drain pan water (105 ml-') of AHU 1and represented a large fraction of the bacterial populationspresent at these sites (74 to 97% and -50% of the totalbacterial plate count, respectively). A yellow-pigmentedBlastobacter species was present to a lesser extent both onthe supply coils and in the drain pan water. Blastobactercells also dominated a large bacterial population on thesupply coils of AHU 2 (Table 2).

It was concluded that a biofilm comprising predominantlythe budding bacterium was present on the coil fins. Thepresence of the biofilm was confirmed by scanning electronmicroscopy of small offcuts of the fin alloy removed from thesupply-side coil bank. A confluent biofilm comprising pre-dominantly large rods (0.6 to 0.8 ,um by 1.4 to 2.8 ,um) wasobserved (Fig. 2). Electron micrographs of the coil fins onthe return side of the cooling coils (data not shown) revealeda surface randomly populated with rodlike cells and fungalspores and mycelia consistent with the diverse heterotrophiccounts obtained for the return coils.Blastobacter cells lost from the biofilm into condensate

would also explain their presence and prevalence in the drainpan water and postcoil air when the coils were functioningand condensate was forming. Viable Blastobacter cells mayalso have been delivered from the coils into the occupiedspace, as evidenced by low numbers present in the supply airsample (4% of the total bacterial count). Other generaidentified from the supply air, particularly Staphylococcusspp., may be shed from the building occupants rather thanfrom the ductwork.Flavobacterium spp., predominantly Flavobacterium od-

oratum, were found in the evaporative condenser sumpwater, on the air intake grill, in the return-fresh air mix, onthe return coils, in the drain pan water, and in the supply airon one sampling occasion. This suggested that the source ofF. odoratum may have been the entrained aerosol from theevaporative condenser sump. Pseudomonas and Bacillusspp. were found at various sites in AH-U 1 but were of variedspecies composition and so unlikely to have a commonsource. Species of Acinetobacter, Cedecea, Arthrobacter,Corynebactenum, and Staphylococcus were found sporadi-cally throughout the AHS. No Legionella spp. were recov-ered from either the drain pan water or evaporative con-denser sump water.

Blastobacter isolates. The Blastobacter colony type wasrepresented by a slowly growing rod that produced small pinkcolonies on R3A in 3 to 4 days and could not be identified bythe Vitek Senior system because of this slow growth. With therudimentary scheme of LeChevallier et al. (18), 33 smallpink-pigmented colony type isolates were identified as Pseu-domonas or Pseudomonas-like, that is, gram-negative, non-fermentative, motile rods. However, phase-contrast micros-copy of these isolates and electron microscopy of negativelystained cells of representatives of these isolates revealed

APPL. ENvIRON. MICROBIOL.

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from

HETEROTROPHIC BACTERIA IN AN AIR-HANDLING SYSTEM 3917

C,,

s C

2._

I~~~~~~~\ I

'-' --O C8C#

o _

-~~~~00

"'-~ 00C -4j

00 -

400:k0-. 00

00 _4~

x-

04

t4-

ONcrlx-

'C

t,-~ 0

0

-A 0

n

-t C

42

1-,-

00~ ~

I- I- -- .- - 0

00 O0

VOL. 58, 1992

(Ar_

0

P.co

CD

00

CO

afD4

'I

00oo

CO

_.

<CO

0' -.

-o_

(A

la a

CO

CD CD3 a

'-1 00

'IC

C COso -o

0- 0CCOv

-0\00 0-0~

x

q. O. 3

-to--0>

-0

-~

0KCtC

t-

t-_c

1--.A

80-

p.

0 0

W' <_

0 _.

0'

CO _.

0~-

U'0eB CO~CLO0 CO

CO

0i*

CD0u_

CD e

CD

0

CO

1.

p-<C O

nCOz

--O<o0.

o .

CO0'00-

00cO00

CL.0

_.

0-

5'CO

5-0'

CO00

CO

0.CO

A

x

00

00

t-\t-

x

0>f-

-+

rnTi

tz

co

0CO

0'-10

0CD(A

0

c0

-0

Ui'COQ0.CD_.

COCD(A

Ct

-A

0 0

-

0C)

C

0

0'

s

CD

Bw

3

0

0

0

co

CO

I

CD

00

..s

P.-

x

4

(-n~

00

-

t-

0

-4

0)-;0Z~0-It03 Z

_ w Pr.p- -

00 w-C

0 0

0

x

1-

I-

- -

00_

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from

3918 HUGENHOLTZ AND FUERST

FIG. 2. Scanning electron micrograph of supply-side coil finbiofilm, showing a dense distribution of bacterial rods embedded inglycocalyx. Bar, 2.0 ,um.

large cells (0.7 to 1.2 ,um by 2.0 to 7.8 ,um), often arranged inrosettes and exhibiting polar buds (Fig. 3). Note that the sizerange of the Blastobacter cells was comparable to the size ofthe large rods observed by electron microscopy in thebiofilm (Fig. 2). Negatively stained cells did not possesscrateriform structures on their cell walls. Thin sections oftwo representative isolates indicated the absence of intra-cellular membranes. Methanol extracts of pigments fromrepresentative strains displayed no peaks consistent withbacteriochlorophylls. These characteristics were consistentwith the genus Blastobacter (see Discussion).

DISCUSSIONIn past microbiological studies of AHSs where bacterial

contamination has been observed, such contamination iscommonly prolific and associated with obvious biofouling ofmoist surface habitats in the system (humidification systems[1, 2, 11, 16, 20], stagnant drain pan waters [16, 20], andcooling coils [13, 20]). In addition, these findings have mostlybeen demonstrated in association with building-related ill-ness. In the present study, we examined an AHS with no

reported building-related illness; in this "healthy" system,these types of gross contamination were not apparent. TheAHUs lacked humidification systems, the drain pans drainedproperly so that stagnant water could not accumulate, andcooling coils were cleaned annually with a high-pressuredetergent spray. Despite the lack of visible signs of microbialcontamination, the AHS harbored significant reservoirs of

bacteria, primarily as a biofilm on the cooling coils. Thebiofilm was invisible to the naked eye, comprised predomi-nantly bacteria belonging to the genus Blastobacter, andgrew on regularly wetted surfaces of the cooling coils (thatis, the supply surface) in both AHUs. The biofilm waspresent on several sampling occasions over the course of ayear with little change in composition despite the annualcleaning of the coils (in August).

Previous descriptions of cooling coil biofilms include amicrobial slime several millimeters thick of unknown com-position reported on the cooling coils in a building demon-strating building-related illness (20) and a biofilm occurringon aluminum cooling coil fins of heat pump equipment withan offensive odor (13). Bacteria isolated and identified fromthe heat pump coil biofilm, like those isolated from thebiofilm in this study, were pseudomonads, Bacillus spp.,coryneforms, and Flavobacterium spp. Bacteria not encoun-tered in the coil fin biofilm of this study but found on the heatpump coil fins included vibrios, azotobacters, Microcyclusspp., Alcaligenes spp., Spirillum spp., and staphylococci.The fact that the biofilm in this study was dominated byBlastobacter cells may be unique to this system, or it may bethat budding bacteria have a competitive advantage overother bacteria in this specialized habitat of the cooling coilsystem. It is unlikely that the Blastobacter strains canoutgrow competitors in the biofilm, since these were themost slowly growing isolates (last to appear as colonies) onthe recovery medium. Dominance is therefore more likelydue to the ability of such bacteria to adhere to the coil finalloy and form the primary colonizing biofilm or to surviveirregular desiccation and temperature variation. The effect ofcontinual desiccation and rehydration on a biofilm warrantsinvestigation.The low residence time of water in the drain pan (due to

washout caused by condensate inflow and drainage outflow)would suggest that the Blastobacter bacteria, being rela-tively slowly growing, would be unable to maintain theirnumbers in the drain pan water through growth. Therefore,Blastobacter bacteria in drain pan water presumably origi-nate from cells continuously lost from the coil biofilm,possibly as swarmer cells shed into the condensate fluid,which runs into the drain pan. Such continuous reinoculationmight explain why other, faster-growing bacteria (for exam-ple, Pseudomonas or Flavobacterium spp.) isolated fromdrain pan water were unable to dominate the drain pan watermicroflora. Cells lost from the coil biofilm into condensate,which is then aerosolized, would also account for the pres-ence of the Blastobacter bacteria in postcoil air and possiblysupply air.The budding bacteria were identified as members of the

genus Blastobacter, demonstrating the distinctive buddingcell morphology characteristic of the genus as defined inBergey's Manual of Systematic Bacteriology (29). Cell ar-rangement in rosettes, a feature of some species in the genus(e.g., B. henrici, B. natatorius, and B. aggregatus), was alsodisplayed by our isolates. Biochemical tests (Gram reactionand oxidase, catalase, and oxidation-fermentation tests)were also consistent with membership in the genus. Inaddition, the absence of intracellular membranes and crater-iform structures distinguished the isolates from methano-trophs and planctomycetes, respectively; the absence ofbacteriochlorophylls distinguished them from the buddingphotosynthetic bacteria in the genus Rhodopseudomonas;and the absence of hyphae distinguished them from Hy-phomicrobium spp. (29).

Blastobacter species have by and large been isolated from

APPL. ENvIRON. MICROBIOL.

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from

HETEROTROPHIC BACTERIA IN AN AIR-HANDLING SYSTEM 3919

FIG. 3. Transmission electron micrograph of a Blastobacter isolate (BF 15) in rosette formation. The cells were negatively stained with1% uranyl acetate in 0.4% sucrose. Arrows indicate buds. Bar, 2.0 ,um.

freshwater habitats such as lakes, ponds, groundwater, anda swimming pool (29), which is consistent with the conden-sate water habitat of the cooling coils and drain pan. The factthat similar budding bacteria have not been reported previ-ously in AHSs may be due to low-nutrient recovery mediaand longer incubation times not being used. Indeed, Blasto-bacter spp., like those of other budding bacterial genera,may be found in many freshwater habitats and may onlyhave been overlooked until now because of difficulties inisolation and recognition of the unusual morphotype (14).The estimation of the numbers of a particular group by

relating the identities of representative isolates to those ofthe same colony type undoubtedly resulted in a loss ofquantitative resolution and an underestimation of diversity.Like any method involving dilution plate counts, the methodis also biased toward numerically dominant members of thebacterial community, although this has the advantage thatmembers that compose significant proportions of the com-munity are highlighted. It also has the advantage of supply-ing isolates for detailed study and resulted in identification ofmembers of an unusual genus, Blastobacter, which wouldnot have been detected without detailed examination of purecultures.

Bacteria previously isolated from an occupied space and ahumidifier and identified as etiologic agents of building-related illness include B. subtilis (17) and Flavobacteriumspp. and their associated endotoxin (23). In the test AHS,Flavobacterium isolates were found to be significant com-

ponents of both the condenser sump and drain pan water.Bacillus isolates, on the other hand, were only present at anyappreciable level on the surface of the supply air registermost remote from AHU 1. Bacteria previously reported inair taken from swine confinement buildings were identified,in order of frequency of detection, as species ofAcinetobac-ter, Alcaligenes, members of the family Enterobacteriaciae,and Pseudomonas (7). In the present study, representativesof these groups, except Alcaligenes spp., were detected inair samples from the AHS.

Bacterial biofilms on AHS surfaces may have severalconsequences. The presence of an established biofilm on thecooling coil fins may represent a significant engineeringproblem in terms of reduced heat transfer rates and theconsequent energy wastage and increased running costs(4-6). Second, biofilm-derived bacteria in the AHS may beimportant in the etiology of building-related illness. Viablebacteria may not need to act as infectious units for theirpresence to be significant for the health of building occu-pants. Lipopolysaccharides (endotoxins) present in the outercell wall membrane of gram-negative bacteria are known tobe a contributing factor in human lung disease and also tocause acute symptoms such as airway constriction caused bymacrophage and neutrophil activation (15, 24, 25, 27). There-fore, the presence of a predominantly gram-negative residentpopulation on surfaces within the system, as in the case ofthe Blastobacter bacteria on cooling coil fins in this study,may be of potential significance to building-related illness

VOL. 58, 1992

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from

3920 HUGENHOLTZ AND FUERST

and its etiology. This is especially so if these bacteria or theirendotoxins are shed into the airstream and delivered to theoccupied space, as appeared to be the case in this study.

ACKNOWLEDGMENTS

We thank Vitek Systems for the use of a Vitek Senior System;Buildings and Grounds Section, University of Queensland, forallowing the study of the test air-handling system; and HelenHosmer for excellent assistance with ultramicrotomy and micro-graph print development. We also thank Ian MacRae for valuableguidance of P.H. during the initial stages of this study.

This work was supported by a grant from Sunny Queen EggFarms Pty. Ltd., Queensland, and by an Australian postgraduateresearch award to P.H.

REFERENCES1. Arnow, P. M., J. N. Fink, D. P. Schlueter, J. J. Barboriak, G.

Mallison, S. I. Said, S. Martin, G. F. Unger, G. T. Scanlon, andV. P. Kurup. 1978. Early detection of hypersensitivity pneumo-nitis in office workers. Am. J. Med. 64:236-242.

2. Banaszak, E. F., W. H. Thiede, and J. N. Fink. 1970. Hyper-sensitivity pneumonitis due to contamination of an air condi-tioner. N. Engl. J. Med. 283:271-276.

3. Brief, R. S., and T. Bernath. 1988. Indoor pollution: guidelinesfor prevention and control of microbiological respiratory haz-ards associated with air conditioning and ventilation systems.Appl. Ind. Hyg. 3:5-10.

4. Characklis, W. G. 1983. A rational approach to problems offouling deposition, p. 1-31. In R. W. Bryers (ed.), Fouling ofheat exchange surfaces. United Engineering Trustees, NewYork.

5. Characklis, W. G. 1990. Microbial fouling, p. 523-584. In W. G.Characklis and K. C. Marshall (ed.), Biofilms. John Wiley &Sons, Inc., New York.

6. Characklis, W. G., M. J. Nimmons, and B. F. Picologlou. 1981.Influence of fouling biofilms on heat transfer. Heat TransferEng. 3:23-37.

7. Clark, S., R. Rylander, and L. Larsson. 1983. Airborne bacteria,endotoxin and fungi in dust in poultry and swine confinementbuildings. Am. Ind. Hyg. Assoc. J. 7:537-541.

8. Clement, E., R. De Koster, and N. V. De Kobra. 1989. Treatmentprogramme for microbiologically contaminated airconditioninginstallations, p. 421-426. In C. J. Bieva, Y. Courtois, and M.Govaerts (ed.), Present and future of indoor air quality. El-selvier Science Publishers B.V., Amsterdam.

9. Edelstein, P. H. 1981. Improved semiselective medium forisolation of Legionella pneumophila from contaminated clinicaland environmental specimens. J. Clin. Microbiol. 14:298-303.

10. Feeley, J. C., R. J. Gibson, G. W. Gorman, N. C. Langford,J. K. Rasheed, D. C. Mackel, and W. B. Baine. 1979. Charcoal-yeast extract agar: primary isolation medium for Legionellapneumophila. J. Clin. Microbiol. 10:437-441.

11. Fink, J. N., E. F. Banaszak, W. H. Thiede, and J. J. Barboriak.1971. Interstitial pneumonitis due to hypersensitivity to anorganismn contaminating a heating system. Ann. Intern. Med.74:80-83.

12. Finnegan, M. J., and A. C. Pickering. 1986. Building relatedillness. Clin. Allergy 16:389-405.

13. Harris, E. F., and C. A. Lee. 1991. Biofilm formation on heatpump equipment, abstr. Q-199, p. 309. Abstr. 91st Annu. Meet.Am. Soc. Microbiol. 1991. American Society for Microbiology,Washington, D.C.

14. Hirsch, P., and M. Muller. 1986. Methods and sources for theenrichment and isolation of budding, nonprosthecate bacteriafrom freshwater. Microb. Ecol. 12:331-341.

15. Hudson, A., K. Kilburn, G. Halprin, and W. McKenzie. 1977.Granulocyte recruitment to airways exposed to endotoxin aero-sol. Am. Rev. Respir. Dis. 115:89-95.

16. Hughes, R. T., and D. M. O'Brien. 1986. Evaluation of buildingventilation systems. Am. Ind. Hyg. Assoc. J. 47:207-213.

17. Johnson, C. L., I. L. Bernstein, J. S. Gallagher, P. F. Bonventre,and S. M. Brooks. 1980. Familial hypersensitivity pneumonitisinduced by Bacillus subtilis. Am. Rev. Respir. Dis. 122:339-348.

18. LeChevallier, M. W., R. J. Seidler, and T. M. Evans. 1980.Enumeration and characterization of standard plate count bac-teria in chlorinated and raw water supplies. Appl. Environ.Microbiol. 40:922-930.

19. Molina, C. 1989. Sick building syndrome-clinical aspects andprevention, p. 15-21. In C. J. Bieva, Y. Courtois, and M.Govaerts (ed.), Present and future of indoor air quality. El-selvier Science Publishers B.V., Amsterdam.

20. Morey, P. R., M. J. Hodgson, W. G. Sorensen, G. J. Kullman,W. W. Rhodes, and G. S. Visvesvara. 1986. Environmentalstudies in moldy office buildings. Am. Soc. Heating Refrig. AirCond. Eng. Trans. 92:399-419.

21. Phillips, E., and P. Nash. 1985. Culture media, p. 1051-1092. InE. H. Lennette, A. Balows, W. J. Hausler, Jr., and H. J.Shadomy (ed.), Manual of clinical microbiology, 4th ed. Amer-ican Society for Microbiology, Washington, D.C.

22. Reasoner, D. J., and E. E. Geldreich. 1985. A new medium forthe enumeration and subculture of bacteria from potable water.Appl. Environ. Microbiol. 49:1-7.

23. Rylander, R., R. Haglind, M. Lundholm, I. Mattsby, and K.Stenqvist. 1978. Endotoxins and the lung: cellular reactions andrisk for disease. Prog. Allergy 33:332-344.

24. Rylander, R., and M.-C. Snella. 1983. Endotoxins and the lung:cellular reactions and risk for disease. Prog. Allergy 33:332-344.

25. Rylander, R., S. Sorensen, H. Goto, K. Yuasa, and S. Tanaka.1989. The importance of endotoxin and glucan for symptoms insick buildings, p. 219-226. In C. J. Bieva, Y. Courtois, and M.Govaerts (ed.), Present and future of indoor air quality. ElsevierScience Publishers B.V., Amsterdam.

26. Shelley, A. 1990. The prevention of microbial contamination inair handling systems. Aust. Refrig. Air Cond. Heating 44(3):30-36.

27. Snella, M.-C., and R. Rylander. 1982. Lung cell reactions afterinhalation of bacterial lipopolysaccharides. Eur. J. Respir. Dis.63:550-557.

28. Standards Association of Australia. 1989. Air handling and watersystems-microbial control. Publication AS 3666-1989. Stan-dards Association of Australia, Sydney.

29. Trotsenko, Y. A., N. V. Doronina, and P. Hirsch. 1989. GenusBlastobacter, p. 1963-1968. In J. T. Staley, M. P. Bryant, N.Pfennig, and J. G. Holt (ed.), Bergey's manual of systematicbacteriology, vol. 3. The Williams & Wilkins Co., Baltimore.

30. Winn, W. C., Jr. 1988. Legionnaires disease: historical perspec-tive. Clin. Microbiol. Rev. 1:60-81.

APPL. ENvIRON. MICROBIOL.

on Novem

ber 10, 2015 by University of Q

ueensland Libraryhttp://aem

.asm.org/

Dow

nloaded from