 

‘1 N1
in." ‘\"~. 1-}
‘IT‘ﬁk‘kt o,{l_. win"

‘«‘-',

131i .. f. .
5 an
.w.

c_
5

A . 5 '_ ‘. __ . ’ ' . -.'4 ... "
‘ -‘ 1>~ _ T . - ._5 '5 “- .A . . .., ' ' ' ’1
\- . . < - '. ‘ A - ‘ ‘ ‘ " 4 1 v. ‘ 2

‘4‘;l.l41“‘r'“
h M" "'51 1
" c 11.453 4:":

"$4 ”5“- -"}r\_\\\

!—~>}L iii“ "‘
.

.-
‘- .‘._:‘
-'-o_

I. v-v 0-0

j: I! :-‘:4 "
1 "v 1%..- 3 ‘9‘
. Elitftg’i. '[u

~1“}“'l‘

r -:

  . 3". A;
"“1359; ’
I'D

ugh-5‘37

’. . 5"
«2-24?

#317

~
f

~17‘v. :.:£;.§E:: \a
1:553} a $1, ;. Leg-,1"
. \‘%1:\’§?‘§§ "1. . ";‘

' '0
I’-

3?“?‘9:

..-
,-..‘

a} 55..

5:"

3:1 ~‘
0. r
;

.3:
“5%.;

‘—
. d‘
' 1:.

‘ﬁﬁx
v ‘2‘; .
.. 0

x.
. ‘2.
.l V”
-;. .2...

- u.-_.x—
-J~——.——A‘

17%.“?
.w.
L..- #-
«y»- ‘-~‘-‘-1
2}"; .
w-
. a

i

L‘
its—«3‘

w—

«9—,

 

'41-.

gr

(1.}:

J

Ill
’1‘.

.-
N‘

é: IIIIII I IIIIII III IIIII II IIIIIIIIIIIIIIIIIIII

5323

This is to certify that the

dissertation entitled

INFLUENCE OF SOIL AGE AND VEGETATIVE COVER ON MICROBIAL
COMMUNITY COMPOSITION: A RIBOSOMAL DNA ANALYSIS OF
HAWAIIAN SOILS

presented by
Klaus R.L. Nﬁsslein

has been accepted towards fulﬁllment
of the requirements for

Ph . D . _ Microbiology
degree in

 

 

 

 

{Jamelgfjowfoﬁﬁoédje

Date May 12,1998

 

MSU is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY

, Mlchlaan State

Unlveralty

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE

DATE DUE

DATE DUE

 

Affil’.
or“: ‘2 4 A’ ~

 

EZW «on

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M chleGS—p.“

 

INFLUENCE OF SOIL AGE AND VEGETATIVE COVER ON
MICROBIAL COMMUNITY COMPOSITION:
A RIBOSOMAL DNA ANALYSIS OF HAWAIIAN SOILS

By

Klaus R. L. Nﬁsslein

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Microbiology

February 1998

Copyright by
KLAUS R. L. NUSSLEIN
1 998

ABSTRACT

The inﬂuence of two environmental factors, soil age and vegetative cover, on the
structure and composition of soil microbial communities was investigated using
Hawaiian soils and DNA-based methods. Hawaiian island soils were expected to have
lower diversity than continental soils because they are young, geographically isolated,
harbor low plant diversity, and the islands are of relatively constant climate. Molecular
methods were used to identify some of the dominant and rare bacterial community
members, and yielded reproducible community proﬁles within sites but distinguishable
between sites. The DNA was ﬁrst fractionated based on its guanine and cytosine
content (G+C), and then rDNA ampliﬁed from a fraction with high biomass (63%
G+C) and a fraction of low biomass (35% G+C). The rDNA clone libraries derived
from these fractions were screened by ampliﬁed rDNA restriction analysis (ARDRA)
to determine phylotype distribution for each soil.

For the 63% G+C fractions the two major changes in community structure were
seen for a chronosequence of soils of ages 200, 2,100, 20,000, and 150,000 years. The
maximum in phylotype diversity at the 20,000 year old site is correlated with maxima
in plant available nitrogen and soil organic matter, while minimum diversity at the

150,000 year old site corresponded to a maximum rate of litter decomposition. The

ﬁndings are consistent with the macro—ecological principle of diversity responses to
nutrient availability. The rDNA method was reproducible within sites, however,
dominant phylotypes were different for each site.

Comparison of the oldest rainforest soil and a soil of an adjacent pasture created
by clear cutting eight decades ago revealed a major shift in the microbial community
structure and composition. The G+C proﬁle showed a shift to organisms with higher
G+C content in the pasture. None of the nucleotide sequences analyzed matched
sequences in the databases, and hence reﬂect novel bacteria.

The two step rDNA analysis used here uncovered new rDNA diversity and
provided evidence that soil microbial diversity is much higher than can be revealed in
eubacterial clone libraries ﬁom total community DNA. This work provides the ﬁrst
evaluation of the response of dominant and rare members within a soil microbial
community to environmental factors, and revealed that vegetation was a particularly
strong factor in shaping microbial community composition. Understanding what
parameters and principles inﬂuence soil microbial community composition may aid

prediction and management of soil biological processes.

This dissertation is dedicated to my parents Hildegard and Rudolf,
and to my wife Christina

ACKNOWLEDGMENTS

I would like to express gratitude to my major professor and research advisor,
James M. Tiedje, for his instruction, guidance and encouragement during my Ph.D
study. I have learned much by his example, and I will remain grateful for all the
opportunities he made possible for me. Special thanks are extended to my advisory
committee, Mike Klug, Eldor Paul, and Loren Snyder for their instruction and
encouragement throughout this research work, for the time and effort spent on reviewing
this dissertation, and for their valuable comments and suggestions.

Special thanks are extended to Mary Ann Bruns, Larry Fomey, Joyce
Wildenthal, Bernard Schroeter, Dave Harris, Tom Schmidt, John Urbance, Jizhong
Zhou, and Dave Odelson for their invaluable help and assistance during this work. I also
thank my fellow students and coworkers for their friendship, support and stimulating
discussions during my years at Michigan State University.

This study was funded by a grant from the National Science Foundation Center
for Microbial Ecology at Michigan State University under Microbial Ecology Grant

NSF DEB #9120006.

vi

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................... ix
LIST OF FIGURES .................................................................................................... xii
LIST OF ABBREVIATIONS .................................................................................... xix
CHAPTER 1
INTRODUCTION AND RATIONALE
Background ....................................................................................................... l
1. Soil microbial communites ...................................................................... l
2. Why Hawaii ............................................................................................ 3
3. Ribosomal RNA as a tool in microbial ecology ...................................... 5
4. Molecular microbial diversity of soils .................................................... 9
5. Potential biases of SSU rDNA-based studies ....................................... l9
Purpose and Goal of this Study ...................................................................... 22
Outline and Thesis Design .............................................................................. 24
CHAPTER 2

CHARACTERIZATION OF THE DOMINANT AND RARE
MEMBERS OF A YOUNG HAWAIIAN SOIL BACTERIAL
COMMUNITY WITH SMALL-SUBUNIT RIBOSOMAL DNA
AMPLIFIED FROM DNA FRACTIONATED ON THE BASIS OF

ITS GUANINE AND CYTOSINE COMPOSITION ............................................... 26
Abstract .......................................................................................................... 27
Introduction .................................................................................................... 27
Materials and Methods ................................................................................... 27
Results ............................................................................................................ 28
Discussion ....................................................................................................... 31
References ....................................................................................................... 32

vii

Table of contents, continued

CHAPTER 3
SOIL BACTERIAL COMMUNITY SHIFT LINKED TO CHANGE
FROM FOREST TO PASTURE VEGETATION IN A TROPICAL

SOIL ............................................................................................................................ 34
Abstract .......................................................................................................... 34
Introduction .................................................................................................... 35
Materials and Methods ................................................................................... 37
Results ............................................................................................................ 41
Discussion ....................................................................................................... 50

CHAPTER 4

SOIL MICROBIAL COMMUNITY DEVELOPMENT AS

INFLUENCED BY SOIL AGE ................................................................................. 53
Abstract .......................................................................................................... 53
Introduction .................................................................................................... 55
Materials and Methods ................................................................................... 57
Results ............................................................................................................ 64
Discussion ....................................................................................................... 88

APPENDICES .......................................................................................................... 100
Appendix A: Statistical comparison of microbial communities by

rarefaction curves ................................................................... 100

Appendix B: Statistical analysis of ARDRA derived rank-
abundance patterns ................................................................. 107
BIBLIOGRAPHY .................................................................................................... 130

viii

LIST OF TABLES

CHAPTER]

Table 1. Characteristics and differences of seven published rDNA
sequenced-based analyses of soil bacterial communities ............................... 12

CHAPTER 2

Table 1. Phylogenetic affiliations based on SSU rDNA genes of
members of a Hawaiian rainforest soil bacterial community ......................... 29

CHAPTER 3

Table 1. Differences in soil properties between the Kohala forest and
pasture soils .................................................................................................. 42

Table 2. Phylogenetic afﬁliations based on SSU rDNA genes of the

three most dominant phylotypes of the two replicates from
the 63% G+C fractions of the rainforest soil (HRS) and the
pasture soil (HPS) ......................................................................................... 46

Table 3. Evolutionary distances among SSU rDNA clones from
dominant members of the 63% G+C soil bacterial community
of Kohala rainforest and pasture soils. The numbers represent
percent SSU rDNA nucleotide sequence similarity ...................................... 49

ix

List of Tables, continued

CHAPTER4

Table l. Site descriptions and soil properties of sampling locations ........................... 59

Table 2. Plate counts, direct counts, cell lysis efﬁciency and DNA
yields for individual samples of ﬁve soils of culturable
heterotrophic bacteria ................................................................................... 65

Table 3. Comparison of extracellular and intracellular DNA extraction
yields ............................................................................................................. 68

Table 4. Number of clones and phylotypes in clone libraries and three
most dominant phylotypes for each soil sample in the
chronosequence. Similar phylotypes are indicated by bold
numbers in identically shaded boxes. Underlined numbers
indicate clones analyzed by nucleotide sequencing (Table 6) ....................... 74

Table 5. Statistical comparison of community patterns over time. After
the test results for the Kruskal-Wallis statistic indicated
nonhomogeneity between communities of the chronosequence
the Multiple Comparisons Procedure was used to determine if
the community structures are signiﬁcantly different from each
other (p<0.05). All clone libraries were grouped by ranks prior
to the test. Details see Appendix B ............................................................. 82

Table 6. Phylogenetic afﬁliations of SSU rDNA sequences from a
chronosequence of Hawaiian rainforest soil (HRS). The
relative abundance of each phylotype is indicated, as well as
the similarity to the closest known species. Only
unambiguously aligned regions were used for the analysis. All
clones are from the 63% G+C fractions for each soil .................................... 85

List of Tables, continued

APPENDIX

Table 1. Contingency table to classify and rank different populations in
categories ..................................................................................................... 109

Table 2. Contingency table for four soil bacterial communities from soils
of different age. Observed counts are classiﬁed into three
ordered categories. The number in parentheses denote the
expected cell frequencies. All data are derived from the 63%
G+C fraction. The data for the two oldest soils (populations
k3 and k4) are contributed by three replicates each .................................... 112

Table 3. Results for the Multiple Comparisons Procedure of four
different soil bacterial communities ............................................................. 1 13

Table 4. Comparison of factors a and b in the regression equation for
rank-abundance distributions of four soils of different age
(63% G+C fraction) .................................................................................... 125

Table 5. Comparison of factors a and b in the geometric regression
equation for rank-abundance distributions of all soils
investigated (63% G+C fraction) ................................................................ 126

LIST OF FIGURES

CHAPTERI

Figure 1. Rooted universal phylogenetic tree showing the three domains,
Archaea, Bacteria, and Eucarya based upon prokaryotic SSU
rRNA sequences (adapted from Woese, 1994) ............................................... 4

Figure 2. Increase of the number of SSU rRNA genes sequenced since 1987 .............. 10

Figure 3. Total soil DNA is fractionated according to its G+C content,
and two single fractions are analyzed for their phylotype rank-
abundance patterns after restriction digestion with two pairs of

restriction enzymes ....................................................................................... 14

Figure 4. Flowchart of the analytical procedure employed in this study
from the extraction of community DNA to the determination
of phylogeny of certain community members. ............................................. 16

Figure 5. The three projects described in this thesis, and their underlying
variables at four soils of a chronosequence in Hawaii ................................... 23

CHAPTER2

Figure 1. Microbial community structure of a young rainforest soil
determined by the G+C content of its DNA. The bacterial

community proﬁle of a mid-Michigan agricultural soil is
included for comparison (Holben and Harris, 1995) ..................................... 29

xii

List of Figures, continued

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Restriction patterns of ampliﬁed SSU rRNA clones in the 63%
G+C fraction after restriction digestion with HaeIII and HhaI. Plasmid
pBR322 digested with HaeIII (marker V) was used as a DNA size marker.. 30

Frequency distribution of SSU rDNA gene phylotypes

(restriction patterns) from the total soil DNA (A), the 35%

G+C fraction (B), and the 63% G+C fraction (C) of a young

Hawaiian soil. The proﬁles are based on results obtained after

digestion with tetrameric restriction endonucleases HaeIII plus

Hhal and Mspl plus RsaI. ............................................................................. 30

Rarefaction curves for the phylotypes found in total soil DNA

and in the 35 and 63% G+C fractions of a Hawaiian soil. The

expected numbers of phylotypes calculated from a random

sample of individuals taken from the total population of

phylotypes are shown on the y-axis ............................................................. 31

DGGE analysis of SSU rDNA fragments (length, 434 bp)
obtained aﬁer PCR ampliﬁcation of the 35 and 63% G+C
fractions of soil DNA and individual SSU rDNA clones from
the clone libraries that were most frequent as determined by
ampliﬁed rDNA restriction analysis. Lane 1, clone HRS-2 of
the 35% G+C fraction; lane 2, 35% G+C fraction; lane 4, 63%
G+C fraction; lane 5, clone HRS-12 of the 63% G+C fraction.
Lane 3 contained a mixture of several bacterial genomic DNAs
as a marker and positive control. The ﬁgure is a negative image
of a silver-stained DGGE separation pattern. The arrows
indicate dominant bands of well represented clones, which
were also found in the respective soil DNA fractions. PCR
products obtained from some strains (lane 5) produced more
than one band due to sequence heterogeneities of 16S rRNA
operons (Muyzer, 1995) ............................................................................... 31

xiii

List of Figures, continued

Figure 6.

Figure 7.

Figure 1.

Figure 2.

Phylogenetic relationships of the most closely related SSU

rDNA clones from the 35% G+C ﬁ'action of the young

Hawaiian soil used. Evolutionary distances were determined

by maximum likelihood analysis. Bar = 0.10 substitution per

base position. env., environmental clone ...................................................... 32

Phylogenetic afﬁliation of the dominant and rare phylotypes

identiﬁed in phylotype abundance distribution proﬁles for the

35 and 63% G+C fractions of Hawaiian soil. The numbers

indicate the designations of the clones that were analyzed (1 ,

HRS-1; 2, HRS-2; etc.). Abbreviations: CIos., Clostridium;

Clos. butyr., Clsotridium butyricum; Agrobact.,

Agrobacterium; Pseudom., Pseudomonas; Rhodosp.,

Rhodospirillum; Acidobact. Subdiv., Acidobacterium

subdivision; Bac.- Lactobac. Subdiv., Bacillus-Lactobacillus

subdivision; Bac., Bacillus ............................................................................. 32

CHAPTER 3

Comparison of G+C proﬁles of DNA extracted from two

adjacent soils in the Kohala area One soil supports a

rainforest and the adjacent soil was converted to pasture 80

years ago. Proﬁles of DNA extracted from replicate samples

at the same site are shown ............................................................................. 43

Frequency distribution of SSU rDNA gene phylotypes from
the 63% G+C fractions (A) and the 35% G+C fractions (B)

from the rainforest and pasture soils. Each graph shows the
median and range 03am) of data of two replicates. Solid dots
(0) are placed above the phylotype which were identical in

pattern and rank for the two replicates. A replicate of the 35%
G+ C pasture sample was not available. ....................................................... 44

xiv

List of Figures, continued

Figure 3.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

An ARDRA proﬁle following double digestion with HaeIII

and Hhal of clones form the 63% fraction of the pasture soil.

Similar patterns can be seen for three groups of clones, each

found in both replicates: lanes 2,5,15; 6,7,8,10,l4,l7, and

9,11,12. Lanes 1 to 8 are from replicate A and lanes 9 to 18

from replicate B. Plasmid pBR322 digested with HaeIII was

used as a size marker (M) ............................................................................. 48

CHAPTER4

Map showing the location of the four soil sites of the
chronosequence on the Big Island of Hawaii, HI, USA ................................ 61

Replication of G+C proﬁles for three soil samples from the

Kohala site (150,000 years). Replicates A and B were

fragments of the same soil aggregate sampled in 1992.

Replicate C was sampled at the same soil site 2 years later. ........................ 67

Distribution of G+C content of DNA extracted from the four

soils of the chronosequence. Insert: Separation of a mixture of

three known standards of genomic bacterial DNA used in the

standardization of this method ...................................................................... 70

Contribution of different %G+C ranges to the community
proﬁles for the four soils in the chronosequence .......................................... 7l

Replication of rank-abundance proﬁles for three soil samples of the mid-
aged Laupahoehoe site (20,000 years). ......................................................... 73

XV

List of Figures, continued

Figure 6. Phylotype rank abundance proﬁles for the 63% G+C fraction
of the four soils of the chronosequence. The proﬁles of the
two youngest soils are single experiments, the proﬁles of the
two oldest soils are averages of three replicates each. Insert:
Ratio between the three best represented clones of the 63%
G+C fractions and the remaining clones within each clone
library along the chronosequence .................................................................. 76

Figure 7. Phylotype rank abundance proﬁles for the 35% G+C fraction
of the four soils of the chronosequence. Insert: Ratio between
the three best represented clones of the 35% G+C ﬁ'action and
the remaining clones within each clone library along the
chronosequence ............................................................................................. 77

Figure 8. Rarefaction curves for the phylotypes found in the 63% G+C
(A), and in the 35% (B) G+C fractions along the
chronosequence. The X-axis depicts the expected number of
phylotypes calculated from a random sample of individuals
taken from the total population of phylotypes. A comparison
with the rarefaction curve for phylotypes from
nonfractionated whole community DNA is shown ...................................... 78

Figure 9. A simpliﬁed diversity index, deﬁned as the ratio of all
phylotypes of contributing clones in a particular clone library
versus the total number of clones in that clone library, is
shown for the four soils. The value of each sample is shown in
each bar as well as the mean and the standard deviation ............................... 79

Figure 10. Plant available nitrogen and phosphorus determined in situ with anion-
exchange resin bags that were buried from 5 to 8 month (data
from Crews et 01., 1995). The data are overlaid with the
diversity index from Figure 9. ....................................................................... 87

xvi

List of Figures, continued

Figure 11.

Figure 12.

Figure 1.

Figure 2.

Figure 3.

Comparison of soil organic matter content and phylotype
diversity for soils of the chronosequence .................................................... 89

Exponential decomposition constants and foliar nitrogen

concentration for leaf litter of the dominant plant

Metrosideros polymorpha. Decomposition constants were

determined over 2 years by in situ decomposition (data from

Crews et al., 1995). F oliar nitrogen concentration is calculated

in % dry mass (data from Vitousek et al., 1995). Net primary

production is shown (data from Herbert, 1995). ........................................ 90

APPENDIX

Sampling plot of the dependence of diversity on the number of

individuals. Region A refers to the curve area of rapid

accumulation of phylotypes associated with increasing number

of clones sampled. As the number of clones added becomes

larger, region B, the clone library has become large enough that

new phylotypes are added at a much slower rate (adapted after

Sanders 1968). ........................................................................................... 101

Example of data entry and output for the rarefaction program
SIM using data from a fraction at the youngest Hawaii site,
Thurston ...................................................................................................... 104

Rarefaction curves calculated by the rarefaction program SIM

for comparison of three related clone libraries. Non-

fractionated DNA (A) is compared with fractions of the same

sample ﬁrst separated by its G+C content for their phylotype

diversity. Since rarefaction curves produce values of E(Sm)

only for integer values of m, the datapoints representing the

rarefaction curve must be identiﬁed. ........................................................... 105

xvii

List of Figures, continued

Figure 4. Clones per phylotype group for the 63% G+C fractions of a
chronosequence of four soils. The two oldest soils are shown
as the mean of three clone library replicates. .............................................. 113

Figure 5. Contribution of expected cell ﬁequencies in the contingency
table (Table 2) to the value of Chi-Square for the 63% G+C
fractions of a chronosequence of four soils (equation [2]) .......................... l 14

Figure 6. Regression trendlines and their equations calculated to ﬁt the
rank-abundance proﬁles for the four soils of the
chronosequence. All graphics describe the 63% G+C fractions
for each respective soil site ......................................................................... 124

xviii

LIST OF ABBREVIATIONS

ARDRA ................................................... Ampliﬁed ribosomal DNA restriction analysis
DGGE .................................................................. Denaturing gradient gel electrophoresis
DTAF ........................................................ 5-(4,6-dichlorotriazin-2-yl)amino ﬂuorescein
G+C .................................................................................................. guanine and cytosine
HRS-1 ............................................................. Hawaiian rainforest soil clone number one
HPS-2 ................................................................. Hawaiian pasture soil clone number two
NPP ............................................................................................. Net primary production
PCR .......................................................................................... Polymerase chain reaction
rDNA ....................................................................................................... ribosomal DNA
Rf ................................................................................................................ refractive index
RFLP ............................................................... restriction fragment length polymorphism
SSU .............................................................................................................. small-subunit

xix

CHAPTER]

INTRODUCTION AND RATIONALE

BACKGROUND

1. Soil microbial communities

Soil microbial communities are some of the most complex, diverse, and important
assemblages of organisms in the biosphere, yet little is known about their structure and
their response to environmental inﬂuences. The complexity of their environment results
in numerous ecological niches. Niche availability and diversity is given among other
factors by mineral composition of the soil, salinity, pH value, nutrient availability,
organic input, temperature, and water content. This sum of basic variables is overlaid
by a high spatial heterogeneity caused by low mixing rates, high surface area, the
formation of macro- and microaggregates, and the inﬂuence of plants and soil animals. In
addition to new niches through the input or the metabolism of organic matter, plants and
animals also create new habitats of increased metabolic activity and complexity such as

the rhizosphere or animal digestive tracks. Finally, seasonal and vegetative changes

superimpose a layer of temporal heterogeneity, resulting in the common notion that no
two soil samples have the same microbial community (Liesack et al., 1997).

As a result of the high niche variability of the soil habitat microbial diversity is
thought to be extremely high, however the extent of that diversity is unknown.
Furthermore, the inﬂuence of soil environmental parameters on this high diversity
remains largely unexplored. Diversity and community structure currently are the focal
points of soil microbial community studies. In this study diversity is deﬁned as the
number of different species present in a soil habitat, or, in molecular terms, the number
of different phylotypes (ﬁngerprint patterns or nucleotide sequences) present in a
habitat. Community structure includes quantitative information on the number of
individuals of different phylotypes or otherwise deﬁned groups.

Analysis of soil microbial diversity presents a major challenge. Molecular
techniques are reinforcing the evidence for extremely high below ground diversity among
the prokaryotes. Torsvik et al. (1990) used reassociation kinetics of soil extracted DNA
to yield an estimate of approximately 4,000 different microbial genomes in one gram of
soil, representing perhaps as many as 13,000 different species (Torsvik et al., 1994).
Diversity data need to be accompanied by data for species or group abundance, since

each strain can vary from one cell per gram of soil to maybe 108.

2. Why Hawaii

Consequently, Hawaii was chosen for the study of soil microbial community
structure and development based on the notion that a young and isolated land mass is
lesser developed than old continental systems, and hence exhibits a natural soil
community but at reduced microbial diversity. The selected sites were some of the most
pristine and simple sites available for addressing questions about whether the particular
factors of age and vegetative type inﬂuence the microbial community structure and
composition.

The Hawaiian Islands are ideal natural laboratories for ecologiml research. A
series of features of the Hawaiian Islands make them particularly suitable for studies of
long term soil and ecosystem development. First, the Hawaiian Islands are the most
geographically isolated archipelago on Earth, yet, they are very accessible. They exhibit
among the youngest soils in the world, with rapid biological development.
Macrospecies richness is at a low level. The few species that colonized have radiated to
occupy an extremely broad range of environments and soils (Carlquist, 1980). Second,
other variables can be held nearly constant. The selected sites have a relatively constant
climate without seasonal changes, and the high mean annual precipitation and
temperature are nearly constant throughout the year. All sites are at about the same
elevation, hence same temperature, are about 1 meter deep, and, with one exception,
support intact rain forest vegetation dominated by the same tree species. The mineral

soil substrate shows no or little chemical variation. The maritime tropical environment

4

reduces the impact of Pleistocene climate change. Glaciation, which would reset soil and
ecosystem development, was precluded. Third, the Hawaiian Islands result from the
movement of the Paciﬁc tectonic plate over a stationary “hot-spot” in the mantle
(Hawaii Scientiﬁc Project Drilling Team, 1996). The distance from the currently active
volcanoes correspond to substrate age. Fourth, the Hawaiian Islands are one of the best
studied areas by ecologists. Much backgrormd information is supplied on these well
selected, representative sites by other research groups working at these sites.

Several critical aspects must not be overlooked. For example, Pleistocene
climatic variations occurred also in Hawaii, if in a dampened manner compared to
continental and temperate areas. Therefore, sites over 14,000 y of age developed under
conditions different from the present (Hotchkiss et al., 1993). Also, subtle effects of
surface erosion, varying exogenous inputs of atmospheric dust from Asia ((Jackson et
al., 1971; Betzer et al., 1988), and the recent introduction of plant and animal species
from outside Hawaii have changed the environment (Cuddihy er al., 1990).
Nevertheless, environmental variation that could affect soils and ecosystems have been

constrained in Hawaii, to an extent that cannot be matched elsewhere.

3. Ribosomal RNA as a tool in microbial ecology

Knowledge of bacterial diversity obtained after more than 100 years of pure
culture study is incomplete, and very few of the total number of microbial species are in
culture (Torsvik er al., 1990; Amann et al., 1995). Until the middle of the last decade
enumeration and identiﬁcation of soil microorganisms had to rely on phenotypic
methods. However, phenotypic methods are restricted to only those bacteria that can
be isolated and cultured. Other, maybe completely unsuspected groups which may be
abundant or very active will not be considered, rendering the emerging picture of the soil
microbial community false. Only a small proportion of all prokaryotes have so far been
cultivated, and the majority of soil bacteria observed microscopically cannot be
cultivated. In addition, the selective enrichment culture has severe limitations as an
approach to study the community composition of naturally occurring microorganisms
(Ward et al., 1992; Amann et al., 1995). Results are not always reproducible due to the
variability of phenotypic properties in relation to culture conditions. Furthermore,
laboratory cultivation introduces serious bias to community analysis (Boivin-Jahns er
al., 1995; Ferris et al., 1996; Rheims et al., 1996), since the bacterial populations
obtained through plating are mainly dependent on the isolation media used (Serheim et
al., 1989), as well as on puriﬁcation and maintenance procedures. Since nutrient-rich
media are used the selection might be biased towards copiotrophic bacteria rather than
dominant community members. In addition. properties of microbes ex situ may differ,

because they lack the interaction among populations or with the natural environment, or

   
  
  
     

 

 

Bacteria Archa ea E ucarya
AMI-Is
Gwen Entanoebae Slime
non-sulur molds
beanie
arryarchuou Fungi
mmmurciu
Gram Cm. Holcomb: Pm"
oeitives Methano-
£233. p 86’1”” bacterium Clinton
Thompmteus 2401;” a: o
Cynnobaeteril Pylodial 730'“ “mum
Flavoblctoﬁn
Tm
Thenmtogalos
W
Diplononuh

Figure 1. Rooted universal phylogenetic tree showing the three domains, Archaea,
Bacteria, and E ucarya based upon prokaryotic SSU rRNA sequences
(adapted from Woese, 1994).

change their physiological state as a result of adaptation to the changed environment.
Consequently, in situ diversity and community structure are unlikely to be represented
by collections of natural isolates.

Alternatively, molecular approaches based on phylogenetic analyses of biological
markers amongst cellular components are capable of better encompassing the diversity
of all well represented members of the entire community. Since the mid-19808, the use
of small-subunit ribosomal ribonucleic acid (SSU rRNA) based techniques has facilitated
a culture-independent approach of investigating microorganisms as they occur in nature
(Olsen et al., 1986; Ward et al., 1992; Amman et al., 1995). The comparison of these
molecular “signature” sequences transformed microbial taxonomy from a pure
identiﬁcation system to an evolutionarily—based framework (Gray et al., 1984; Woese,
1987; Olsen et al., 1994). Based on these studies all forms of life are separated into
three major evolutionary lines, the three so-called domains: Bacteria, Archaea, and
Eucarya (Fig. 1; Fox et al., 1980; Woese et al., 1990). Ribosomal RNA and the
corresponding genes (rm) are now widely used as powerful evolutionary and
investigative biomarkers for the following reasons (Olsen et al., 1986): (i) Ribosomal
RNAs are essential to protein synthesis, and therefore are ubiquitous to all organisms,
and structurally and functionally conserved ; (ii) ribosomal RNAs are readily isolated
and identiﬁed, (iii) they contain variable and highly conserved regions in both primary
and secondary structure, (iv) and they appear to change in sequence very slowly, and
they do not exhibit horizontal gene transfer found with many other prokayotic genes;

therefore relationships between rRNAs reﬂect evolutionary relationships. These traits

make rRNAs not only the most widely used biomarker, but also a powerful tool for
microbial ecology studies, particularly for complex terrestrial environments such as the
soil with enormous and undiscovered diversity.

Amongst a variety of cellular biomarkers studied, the SSU rRNA gene provides
certain aspects of information that makes it an extremely versatile tool and the best
culture-independent biomarker to study microorganisms. Each SSU rRNA gene contains
highly conserved regions found among all living organisms as well as unique variable
regions and hence diagnostic to certain organisms or related groups. Furthermore, the
primary structure of the approximately 1,500 base SSU rRNA gene allows the inference
of phylogenies based on comparative sequence analysis. By estimating the phylogenetic
relatedness to known microorganisms based on the homology of the gene sequence, the
closest affiliation of a newly isolated or molecularly detected microorganism can be
established. In combination with a large and growing SSU rRNA database,
microorganisms can be sorted according to their phylogenetic affiliation, and conversely,
gene probes can be constructed at different levels of speciﬁcity. To date, over 6,000
SSU rRNA sequences have been made available for comparison (Maidak et al., 1996).
As these databases rapidly expand, they constantly improve the process of matching

new sequences to known microorganisms.

4. Molecular microbial diversity of soils

Since the development and widespread application of the polymerase chain
reaction (PCR; Saiki et al., 1988), rRNA sequences can directly be obtained from lysed
cells, which has contributed to the exponential increase in known prokaryotic SSU
rRNA sequences in recent years (Fig. 2). Results from molecular ecological studies
within the last seven years from marine (Giovannoni et al., 1990; Schmidt et al., 1991;
Moyer et al., 1994), thermophilic (Weller et al., 1991; Ferris et al., 1996), terrestrial
(Liesack and Stackebrandt, 1992; Bomeman et al., 1996) environments and on
symbionts (Amann et al., 1991) documents the success of this strategy.

Several studies explored the microbial diversity of soil bacterial communities by
sequencing SSU rDNA clone libraries (Liesack et al., 1992; Ueda et al., 1995; Lee et al.,
1996; Bomeman et al., 1996, 1997; Kuske et al., 1997; Zhou et al., 1997; Ludwig et
al.,1997). In all cases, libraries were prepared by polymerase chain reaction
ampliﬁcation (PCR) of rRNA genes directly from total soil genomic DNA. These
studies reveal extensive diversity of rRNA operons, and many examples of novel
sequences that are only distantly related to those known from cultivated species.
Differences between these studies are summarized in Table 1. None of the studies
reported duplicate sequences in the analyzed clone libraries, which supports the notion
of high diversity. While all the studies analyzed diversity of a particular soil site they

cannot readily be compared, because they either use PCR primer sets of unique

10

 

6000

4000

2000

SSU rRNA sequences

 

 

 

 

o l l l
1988 1990 1992 1994 1996

Years

Figure 2. Increase of the number of SSU rRNA genes sequenced since 1987

11

speciﬁcity (Liesack et al., 1992; Rheims et al., 1996) or they analyze sequences in
different portions of the SSU rDNA.

The studies of soil community structure found to date tend to concentrate on the
relatively few most abundant species, but do not explore the rich variety of organisms
anticipated to occur in smaller numbers. These microorganisms may be important in
acting as a reservoir of different physiological types, which may respond by population
size increase following changes in the habitat. This would also imply a degree of builtin
redundancy in the microbial community. Some organisms likely are also remainders of
past conditions. Knowledge of the composition of this reservoir will allow us to
distinguish between a change in only the dominant community members and a shift of
the whole community. In the present study a different, more sensitive path to
investigate bacterial diversity of the soil was used, namely the application of a guanine
and cytosine (G+C) based fractionation of total soil DNA, and the use of only one G+C
fraction out of approximately 80 as the source of community DNA for SSU rDNA
studies (Fig.3). Since an underlying community structure was not revealed by analyzing
the total eubacterial rDNA, as was shown in this and all of the above mentioned SSU
rDNA library studies, total soil extracted DNA was ﬁrst fractionated by its G+C
content using bis-benzimidazole and equilibrium centrifugation (Holben et al., 1994).
Thereafter, bacterial rDNA was ampliﬁed by PCR using universal eubacterial primers
from a fraction with high biomass (63% G+C), and a fraction of low biomass (35%
G+C). The rDNA clone libraries were screened by ampliﬁed ribosomal DNA restriction

analysis (ARDRA) to determine phylotype distribution. Similar restriction patterns

12

 

Gemmv mow—Nos: .962

 

ascewémmnnoiu
82 as.» 238 :50 «a: 3.225%
:3 he :anEom ataomnooaoi 3. mo? .aﬂozabnn 853m 55.0 .<.m.D 68:03.3?
@5me Swag: .902 2:0on
03— mo>Emom EEO 3:33 EEO-.m
:3 a. £523 «tangoeoi Na 32 ﬁﬁo>EDJm won Eon EaESD .EEEO
$3 .832: E62
Bonnie—m .<. m . D
33 moaooxﬁoaocma M83 Embzcbnm £232me
:3 3 03 38030205 5 mwc . EmB>EDJm ecu—mmﬁw E83
«852
.8636 5:3 520
32 anew 0338 EEO ~50: EEo>EDJm
:3 3 «R5 sundownoouoi E "52 _ 13.63: D-.m 22.. 5238 :82 55:3—
mmE “Swag: _o>oZ
. E as ﬁcﬁnoxoﬁm mouooonaocm—m 2,38%
No2 meoozﬁoczo’x ”dumm— mooonEobm -.m 25.59:
do 3 x885 «58308058 m: "Km Embzcbnm 7.28 22% £2323 .ocgmtm
v new:
958 3.5888 oath—«5s =2.on— ucEtm EoEcemEo
855.3% «88 .553 8?: 8:20 8.55 <73.— mS zom 5:82 oEqmquoO

 

BEE—EEO... Ethane: :8 no 89:53 33968533 <29— uozmznsq :33 .«o mcococumm€ Ea 89:88.85 .— 93:.

13

 

 

Qwvov momma: 352 20m 526

«insomnia—m 3:8 256—?

moaooxesoami $35.3 :o
33 33% 0353 EEO mmmE Baez—53m 3.3883 .<. W:
:3 B 8.3M «EQQBOQEE ow mum Embzcbum Bﬁcshéobq .285 E0552

3.83

SE Mmmm _ .mﬁo>ED-.m 3:62:
:3 3 E263 <\Z 3: EN .mm5>_:D-.m bﬁococuw 32:80 ﬂap—=2

383055
82 3:8» 9368 EEO mav— EEo>EDJm «imam arena
:3 3 BEN «5830805 mv mum Embasbum 89:: 283 .2230: «Ex—0M
QEW—v mownocz _o>oZ

m:o.w-~w~:ao§0 C33 9525

BSBBQE 820 Saw

mououxﬁoaocai 223m 23

32 33% 92mg EEO MNoE Embzcbum .5885“.
:3 x. 58255 «rayoanoeoi oo— momm Emb>EDum Eoﬁob zuﬁm .maEEomSam

u tom:

258 2.5868 Bras .5282 SEE EuE:8_>:o

8:80qu “8.: :98. 5.52 3:20 SETS <72. m3 =om cognac. oEQSwooO

 

€35.23 ._ 23

r ° + I 1
<
z
o
g a
‘5
l.—
0
DNA extracted °\
from soil
L
A
1?
7g
.2 63o/oG+C
0

SSU rDNA

14

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

per pattern

 

12345678910 15 1)..”656667 12345678910 15 Z)

Restriction pattern (Phylotype)

Figure 3. Total soil DNA is fractionated according to its G+C content,
and two single fractions are analyzed for their phylotype rank-abundance
patterns after restriction digestion with two pairs of restriction enzymes

15

were grouped into phylotypes by computer assisted cluster analysis to create rank-
abundance plots (Fig. 4). Only the nucleotide sequences of the three most dominant and
some of the rare clones were determined, and their phylogenetic affiliation identiﬁed by
comparative sequence analysis. The term “rare” is used here relative to dominant
community members, although we cannot speak of rarity in an ecosystem at a
sensitivity of 2104 cells per phylotype.

In addition to lowering diversity to a more manageable level, the G+C-based
fractionation of total soil DNA yields community proﬁles speciﬁc for each soil site.
Evidence from 3H-labeling experiments and G+C based fractionation of soil community
DNA indicates that activity and growth are not conﬁned to a small fraction of the
bacterial biomass but are widespread amongst the genera represented in the proﬁle
(Harris, 1994). Consequently, the two step rDNA analysis used here uncovered more
diversity than can be detected by the direct rDNA analysis of total community DNA.
The G+C separation step is also a means to detect less dominant organisms in a
community.

Since the structure of microbial communities in soil is made up of a high level of
prokaryotic diversity, sequencing large numbers of clones would be too costly and time-
consuming. The restriction analysis of ampliﬁed rDNA (ARDRA) is a tool for the fast
and reliable SSU-rDNA-based differentiation of a wide diversity of microorganisms
without sequencing. The generation of restriction patterns as ﬁngerprints provides a
means for the stepwise exclusion of redundant clones, or for the quantiﬁcation of

repeated patterns. Internet services exist (Herrnjakob, 1997; Kim et al., 1997) that

16

Figure 4. Flowchart of the analytical procedure employed in this study: Soil DNA

extraction, fractionation and bacterial community ﬁngerprinting of ampliﬁed

l6S ribosomal DNA gene sequences and restriction endonuclease analysis
(ARDRA) of a clone library

DNA EXTRACTION
AND PURIFICATION

 

 

 

<
z
D a
3 |
°\° |
0
N 40 50 no 70
%G+C
Forward Reverse

 
 
  

 

 

soil, sediment, water,
gut liquid, faeccs, etc.

mechanical, high
temperature, and
detergents

E . . I 1 l 11
DNA fractionation based
on G+C content of genomic
DNA; fractions of 35% and
63% G+C are indicated

BEBJ;

PCR amplification of
16S rRNA from soil DNA
with universal eubacterial
primers (#)

is a mixture
of copies of soil bacterial
16S rDNA fragments

Ligationgf 16$ rDNA

pieces into vector pCR ll

of
vector pCR/l containing
the 168 rDNA insert into
E. coli ToplO' cells

white colonies indicate
the vector contains a
168 rDNA fragment

Figure 4. (continued)

Clones/ pattern

 

 

Restriction Pattern (Phylotypes)

HPS-32

 

01..

m2:

PCR amplification of
each clone 16S rDNA
insert from whole cells

Digestion with two pairs of

restriction endonucleases

of
restriction length
polymorphisms

WW
analysis“ similarities
between restriction patterns
(phylotypes)

Group identical
restriction patterns and
sort by abundance into
rank-abundance pattern

19

compare restriction patterns against a database of theoretical restriction patterns
generated form a 16S rDNA database (e.g. RPD database; Maidak et al., 1996). These
programs return a list of microorganisms whose computed theoretical restriction

patterns match the laboratory derived patterns.

5. Potential biases of SSU rDNA-based studies

The constructed libraries do not always reﬂect the existing total microbial
community, because, as with every method, bias might occur during its construction.
The intial step, sampling, could be biased. The clone library might not be representative
of the most successful community in this soil due to the potentially patchy distribution
of populations in clumps and aggregates in the soil sample. The mosticritical step in the
construction of a clone library is the extraction of nucleic acids from the ecosystem.
Isolated DNA should reﬂect the existing genetic diversity. But microbial cells do not
lyse equally well. Especially Gram positive cells might be underrepresentated in the
clone library due to their resistence to lysis. Some DNA might get lost during
puriﬁcation treatment. Since previous research suggests (Holben, 1994) that direct lysis
recovers a more representative fraction of the genetic diversity than the alternative cell
extraction technique, this approach was applied in this study.

Another potential source of bias is the PCR reaction, where certain SSU rRNA

genes might preferentially be ampliﬁed (Amann et al., 1995). The primer sequences

20

used in the PCR-mediated ampliﬁcation of SSU rDNA should cover the phylogenetic
diversity present in the sample, but primers can only be as good as the databases on
which they are based. Our databases, however, are focused on economically important
strains, i. e. medical isolates, food and industry strains, and do not reﬂect the high
diversity of terrestrial and other environmental samples.

A third potential source of bias is the PCR coampliﬁcation of mixed genomes
which leads to the formation of chimeric SSU rDNA molecules (Liesack et al., 1991;
Wang et al., 1997). PCR-induced chimeras are formed by annealing of partial-length
fragments of different SSU rDNA genes via highly conserved regions. The following
primer independent elongation phase then creates full ﬁagments, thereby leading to
reports of nonexisting organisms.

A fourth factor that could inﬂuence selective recovery is copy number variation
of the rrn gene between bacterial species which could result in a quantitative
misrepresentation of the community proﬁle. This renders clone numbers of low
abundance per phylotype unreliable for quantitative evaluation. Other, less important
sources of bias in DNA sequence-based analyses of natural communities have been
discussed extensively elsewhere (Ward et al., 1992; Stackebrandt et al., 1993; Liesack et
aL,1997)

In this study the following seven measures were implemented to reduce potential
bias formation: (i) Culturing biases were omitted by using a molecular approach. (ii)
Using the direct lysis method we were capable of extracting high yields of non-sheared

DNA. The fact that Gram negative as well as many Gram positive organisms were

21

detected supports the notion that at least some of the reputed lysis resistant bacteria are
not underrepresented in the extracted DNA. The high molecular weight of the extracted
DNA reduces the probability of chimera formation. (iii) Since primers for the PCR-
based ampliﬁcation of SSU rDNA were universal for conserved regions throughout the
entire (known) bacterial domain in the RDP database it is inferred that the mixture of
ampliﬁed SSU rDNA fragments is as diverse as the genomic DNA extracted. (iv) PCR
products were cloned directly in the vector to avoid the introduction of bias from
uneven vector to insert ratios needed in other cloning techniques. (v) Only 22 to 25 PCR
cycles were used to stay within the quantiﬁable linear range of the ampliﬁcation reaction
series (Suzuki and Giovannoni, 1996). (vi) Soil DNA templates were denatured
immediately before adding them into the PCR reaction assay to free the target DNA
from potential contamination with persistent humic compounds, and to optimally
expose all target sites.

Despite the care taken in the design of this study, there are still method
limitations and biases inherent to each step in community analysis that will inﬂuence the
results and conclusions. Nevertheless, this study used currently available methods as
critically as possible with the present knowledge to minimize the introduction of

potential biases.

22

PURPOSE AND GOAL OF THIS STUDY

This dissertation focuses on the analysis of microbial community composition in a
sequence of comparable Hawaiian soils except for differences of age and different
vegetative cover (Fig. 5). The goals of the present study were to (i) determine how a
major subsample of the soil microbial community changes with soil age, and (ii) to
assess the inﬂuence of environmental parameters soil age and vegetation on species
diversity and dominance structure. Nucleic acid based techniques were used to achieve
these goals. The G+C based fractionation of soil extracted DNA yielded patterns of the
entire soil microbial community that can be compared, and allowed the selective analysis
of soil community subsets. SSU rRNA genes were used as interpreters to distinguish
community diversity and community structure, and to describe dominant and rare
community members under the inﬂuence of soil age and vegetative cover. This study
does not attempt the complete comprehensive representation of community structure in
a soil, but rather uses around one hundred of the most successful members to be
compared between soils. The degree to which environmental parameters affect soil
microbial community structure has implications for both our understanding of soil

ecosystem structure and our management of soil as an ecosystem resource.

23

 

an 83:
23mg SEEM

€265 E 35538550 a mo m=8 Bow 3 moaning 33595 :2: was £35 £5 E costume—o $8.82“ 025 BE. .m 95»:—

 

 

 

 

 

 

 

 

 

: 83.2 i 25.3 i .53 i can
520..“ 2.35— oosooganaww— «.29 5525:. A” WEE.
2 2 .. amigo
22:58.» 5722—20:
m 15.25 N «Ego

24

OUTLINE AND THESIS DESIGN

This dissertation is divided into ﬁve chapters. This introduction constitutes the
ﬁrst chapter. Chapter 2 describes the analysis of a young and geographically isolated
soil for its dominant and rare community members. This study advances the use of
G+C based fractionation of the total soil extracted DNA. By concentrating the
community structure analysis to two fractions, one within an area of low biomass the
other in the G+C range typically high in biomass, the level of sensitivity could be
lowered to determine dominance structures within subsets of the soil community. The
dominant biomass reﬂected by the 63% G+C fraction contained several dominant
phylotypes, while the community members that were less successful (35% G+C
fraction) did not show dominance but a very high diversity of phylotypes. Nucleotide
sequence analysis revealed taxa of the groups expected for the particular G+C contents,
and important validation for the method. Dominant phylotypes in the 63% fraction
were of the Pseudomonas, Rhizobium-Agrobacterium, and Rhodospirillum assemblages,
while all clones sequenced from the 35% fraction were afﬁliated with several
Clostridium assemblages. The two step rDNA analysis used here uncovered more
diversity than can be detected by the direct rDNA analysis of total community DNA.
The G+C separation step is also a means to detect less dominant organisms in a
community. Chapter 2 represents a manuscript published in Applied and
Environmental Microbiology (1998) under the authorship of K. Niisslein and J. M.
Tiedje.

Chapter 3 is a comparison of two adjacent soils with the same long term history

but which now differ in their vegetative cover. The reliability of the G+C fractionation

25

technique could be shown through statistical comparisons of repetitive experiments. A
strong shift in soil bacterial community composition was related to change in vegetation
cover. Sequence analysis of the three most dominant members revealed that their taxa
changed from F ibrobacter and Syntrophomonas assemblages to Burkholderia and
Rhizobium-Agrobacterium assemblages. The rare community members overlapped by
only 5% as determined by ARDRA analysis. This work represents the ﬁrst evaluation
of the response of dominant and rare members of a soil microbial community structure
to a major environmental factor, plant type. Chapter 3 is currently a manuscript in
preparation for submission as a note to Applied and Environmental Microbiology under
the authorship of K. Niisslein and J. M. Tiedje.

Chapter 4 constitutes a study of the inﬂuence of soil age on the community
structure of a chronoseries of four Hawaiian rainforest soils ranging in age of parent
material from 200 y to 150,000 y. The analysis of the 63% G+C fractions
demonstrated that while dominance structures intensiﬁed with soil age, bacterial
diversity decreased. The most abundant phylotype at the oldest soil comprised
45(:t2)% of the bacterial clone library. Microbial diversity was found maximal at an
intermediate soil age, which correlated with extreme values in a variety of soil physical
and chemical parameters. The reproducibility of the technique to screen and analyze
clone libraries by their ARDRA pattern was shown by comparison of the identity of
dominant phylotypes in replicate samples. Chapter 4 is currently a manuscript in
preparation for submission to Ecology under the authorship of K. Nﬁsslein and J. M.

Tiedje.

CHAPTER2

CHARACTERIZATION OF THE DOMINANT AND RARE MEMBERS OF A
YOUNG HAWAIIAN SOIL BACTERIAL COMNIUNITY WITH SMALL-
SUBUNIT RIBOSOMAL DNA AMPLIFIED FROM DNA FRACTIONATED ON
THE BASIS OF ITS GUANINE AND CYTOSINE COMPOSITION

26

27

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 1998. p. 1283-1289

0099-2240/'98/$04.00+0
Copyright 0 1998. American Society for Microbiology

Vol. 64, No. 4

Characterization of the Dominant and Rare Members of a
Young Hawaiian Soil Bacterial Community with Small-Subunit

Ribosomal DNA Ampliﬁed from DNA Fractionated on the
Basis of Its Guanine and Cytosine Composition

KLAUS NUSSLEIN‘ AND JAMES M. 'I'IEDJE"2‘

Center for Microbial Ecology.1 and Department of Crop and Soil Sciences,2 Michigan State University,
East Lansing, Michigan 488244325

Received 18 August 1997/Accepted 12 January 1998

The small-subunit ribosomal DNA (rDNA) diversity was found to be very high in a Hawaiian soil community
that might be expected to have lower diversity than the communities in continental soils because the Hawaiian
soil is geographically isolated and only 200 years old, is subjected to a constant climate, and harbors low plant
diversity. Since an underlying community structure could not be revealed by analyzing the total eubacterial
rDNA, we ﬁrst fractionated the DNA on the basis of guanine-plus-cytosine (G+C) content by using bis-
benzimidazole and equilibrium centrifugation and then analyzed the bacterial rDNA ampliﬁed from a fraction
with a high biomass (63% G+C fraction) and a fraction with a low biomass (35% G+C fraction). The rDNA
clone libraries were screened by ampliﬁed rDNA restriction analysis to determine phylotype distribution. The
dominant biomass reﬂected by the 63% G+C fraction contained several dominant phylotypes, while the
community members that were less successful (35% G+C fraction) did not show dominance but there was a
very high diversity of phylotypes. Nucleotide sequence analysis revealed taxa belonging to the groups expected
for the G+C contents used. The dominant phylotypes in the 63% G+C fraction were members of the
Psardomonas, Rhizobium-Agrobacterium, and Rhodorpirillum assemblages, while all of the clones sequenced
from the 35% G+C fraction were aﬁliated with several Closrrwium assemblages. The two-step rDNA analysis
used here uncovered more diversity than can be detected by direct rDNA analysis of total community DNA. The

G+C separation step is also a way to detect some of the less dominant organisms in a community.

 

Soil microbial communities remain some of the most difficult
communities to characterize due to their extreme phenotypic
and genotypic diversity. Estimates of the genotypic diversity in
these communities based on DNA renaturation experiments
suggest that there are 4 x 103 to 7 X 103 different genome
equivalents per g of soil (36), which, if extrapolated to species
diversity, suggests that there are perhaps 10’ or even more
Species per g of soil. Data from cultureobased methods also
suggest that there is high microbial diversity in soil, but these
methods are extremely biased (25, 32) and recover less than
1% of the viable community (3, 20. 36, 39). Molecular ap-
proaches in which rRNA sequences are used to determine the
composition of natural communities identify more of the entire
community. While these approaches also suffer from some
biases and lack resolution at the species level, previous rRNA
characterizations have conﬁrmed that there is a high level of
bacterial diversity in soil communities (4, 20, 35, 38).

To ask meaningful questions about soil community compo-
sition, a more manageable level of diversity (lower diversity) is
needed. We sought to study a community with lower diversity
by focusing on a geographically isolated, young soil, namely,
soil formed from volcanic ash deposited 200 years ago on the
island of Hawaii. Due to the geographic isolation of the Ha-
waiian Islands, the diversity of the native fauna and ﬂora is low
(6, 37). Furthermore, we used the site studied because it ex-
periences a constant annual climate, which could also lessen

‘ Corresponding author. Mailing address: Center for Microbial
Ecology, Plant and Soil Sciences Building, Michigan State University,
East Lansing, MI 48824-1325. Phone: (517) 353-9021. Fax: (517) 353-
2917. E-mail: tiedjej@pilot.msu.edu.

selection for diversity. The low level of diversity in the ﬂora and
fauna has made the Hawaiian Islands an attractive site for
studies on radiation of species and invasion of alien species. If
the soil bacterial community is also less diverse for the same
reasons, not only would it be less complex to analyze, but it
would allow questions about soil community development to
be addressed. An additional advantage of the site selected was
the large amount of previously collected data on plant compo-
sition, ecosystem processes, and soil characteristics which was
available (6, 37).

We analyzed the soil bacterial community diversity of this
ZOO-year-old site by performing an ampliﬁed ribosomal DNA
(rDNA) restriction analysis of a PCR-ampliﬁed rDNA clone
library. The initial analysis revealed a diversity too great to be
captured in a reasonable number of clones analyzed (80 to 90
clones per soil sample). To lower the soil bacterial diversity to
a manageable level, we fractionated the soil DNA on the basis
of G+C content as described by I-Iolben and Harris (12) and

(analyzed rDNA clones in two discrete fractions, a fraction with

1283

a large amount of DNA (the 63% G+C fraction) and a frac-
tion with a small amount of DNA (the 35% G+C fraction).

MATERIALS AND METHODS

Soil origin and soil sampling. Soil was collected from an undisturbed montane
rainforest on the island of Hawaii near Thurston Lava Tube on Kilauea Volcano.
within Volcano National Park (19‘25'N, 155°15'W). The site is dominated by the
native tree species Metrosideros polymorpha (which accounts for 91% of the
cover) and harbors a total of 34 vascular plant species (6). Several tree ferns form
the dominant native understory. which also includes Cirobium spp., Coprosoma
spp.. and Vaccinium claya'num. The site is fenced to keep feral pigs (Sur saofa)
out of the park. ‘Ihe sampling site is located at an elevation of 1.200 m. has a
mean annual air temperature of 16°C with very little variation (8). and has a
mean annual rainfall of 2.500 mm (10) which is well distributed throughout the

1284 NUSSLEIN AND TIEDJE

year due to the relatively constant northeast trade winds (5). The sorl is a Hydric
Dystrandcpt developed on several tcphra (volcanic ash) depositions ranging in
age from 200 to 400 years and is approximately 38 cm deep (31). We rcmovcd the
litter and the ﬁrst 1 cm of soil and sampled 7.5 cm of the upper layer. which was
dcposrted 200 years ago. The samples were collected on the perimeter of thc
Vitousek group’s main study site (6). Duplicate soul samples (750 g each) were
packaged on site in sterile polypropylene bags and immediately put on ice. The
next day they were placed in dry ice coolers and shipped by express mail to
Michigan. where they were stored at -20°C. The soil moisture content was
determmcd by drying soil overnight at 100°C. Soil mechanical and chemical
analyses were done at the Soil Analysis Laboratory. Michigan State University.
by using the methods described by Peck et al. (27).

DNA extraction and puriﬁcation. Soil microbial DNA was extracted from 10 g
of soil by the direct lysis method of Holbcn (11), except that EDTA was not
included to reduce cocxtraction of humic compounds, a low shaker speed was
used to prevent extensive DNA shearing. the shaking time was extended to 45
min. and the phosphate concentration in the lysis butler was adjusted to 100 mM
to overcome the high phosphate absorption capacity of the young minerals. The
subsequent DNA puriﬁcation was also modiﬁed to include agarosc gel puriﬁca-
tion (0.4% agarose) and a single Microcon-lOO microcolumn (Amicon Corp.
Beverly. Mass.) passage of the cxcrsed and melted gel piccc. followed by repeated
washing steps. The extraction efﬁciency was dctcrmincd by comparing the
amount of DNA extracted with the amount of DNA expected. as calculated from
the diﬁcrence between the direct microscopic counts of bacterial cells before and
after lysis. Bacterial cells were counted directly by computer-aided microscopic
counting procedures (43). For consistcnq. all counts were obtained by a single
investigator.

DNA was quantiﬁed by ﬂuorometry (18) with a model TK 100 ﬁuorometer
(Hocfer Scientiﬁc Instruments. San Francisco. Calif.) by using the extended assay
protocol of the manufacturer. Five replicates were used to estimate DNA yields.
Known amounts of lambda phage DNA (Bochrrngcr Mannheim. Indianapolis,
Ind.) were used for all calibrations. The ﬂuorescence intensity of DNA was also
estimated based on the relative intensities in agarose gels of PCR ampliﬁcation
products and restriction digests of known mass.

G+C fractionation technique. DNA fragments were separated on the basis of
G+C content by the procedure of Holben and Harris (12). Brieﬂy, DNA was
mixed with bisbenzimidazolc (Hoechst reagent no. 33258). which binds to ad-
enine and thymidine and changes the buoyant density of DNA in proportion to
its G+C content (40). A gradient of G+C concentrations was then established by
equilibrium density gradient ultracentrifugation. and 0.2-ml fractions were col-
lected with a fraction collector. The DNA in each fraction was quantiﬁed by
spectrophotometry. and its G+C content was established by using a standard
curve relating G+C content to density. which was measured with a Bausch ‘61
Lamb refractometcr. To make PCR ampliﬁcation possible. bis-benzimide and
CsCl were removed from DNA fractions by ﬁve repeated extractions in C50-
saturated isopropanol. followed by spin column chromatography (Wizard PCR
Preps: Promega. Madison. Wis.) with two washing steps. A260 was determined
before and after puriﬁcation to monitor for potential losses of DNA during the
puriﬁcation procedure.

PCR ampliﬁcation of SSU rRNA genes from soil DNA. Small-subunit (SSU)
rRNA genes were PCR ampliﬁed from puriﬁed soil DNA by using eubacterium-
speciﬁc primers 1D] and rP2 of Weisburg et al. (41). PCR were performed with
Taq DNA polymerase (Bochringcr Mannheim) by using the manufacturer’s
protocol, an additional 400 ng of bow'ne serum albumin (Sigma Chemical Co.. St.
Imus. Mo.) per pl. and a model 9600 thermal cycler (Perkin-Elmer Cetus.
Norwalk. Conn). The protocol used consisted of an initial denaturation step
(94°C for 130 s) followed by 25 cycles consisting of denaturation at 94‘C for 60 s,
primer annealing at 55°C for 30 s. and elongation at 72‘C for 120 s plus an
additional 7-min cycle to ﬁnalize the chain reaction. Negative controls without
added DNA. as well as positive controls with pure culture genomic DNA. were
included in all PCR. Aliquots (3 ul) of the ampliﬁed products were separated in
a 0.9% agarose gel by electrophoresis in 1x TAE buffer (22), the gel was stained
with cthidium bromide (500 ng/ul), and the bands were visualized by UV exci-
tation. The PCR products were stored at -20‘C.

To ensure that only soil bacterial rRNA genes were ampliﬁed, the followmg
font quality control steps were used: primer purity was established by high-
perforrnance liquid chromatography; oligonucleotide primers were prepared
fresh from lyophilized stocks for each use; preampliﬁcation heating was used to
maximize PCR sensitivity and Speciﬁcity (7); and the bovine serum albumin stock
solution and the commercial Taq DNA polymerase were tested for potential
contamination with bacterial nucleic acids (29).

Analysis of SSU rDNA clone library. The concentrations of PCR-ampliﬁed
SSU rRNA genes were determined by comparing the ﬂuorescent-band intensi-
ties on agarosc gels to the fluorescent-band intensities of known concentrations
of standard lambda DNA. Prior to cloning. the ampliﬁed SSU rDNA fragments
were puriﬁed by Spin column chromatography (Wizard PCR Miniprep; Pro-
mega). An equimolar amount of ampliﬁed PCR products was ligated to the
vector pCR 11 (lnvitrogen. San Diego. Calif). Ligation and transformation into
Escherichia coli Top-IOF' competent cells were carried out according to the
manufacturer's protocol. A primer pair speciﬁcally designed to complement the
polylinker of the vector pCR ll (44) was used to amplify plasmid inserts directly
from the transformant cells for SSU rDNA gene screening. The following two

28

APPL. Ewtnon. MICROBIOL.

types of negative controls were used in the ampliﬁcation of clone inserts: controls
without target DNA added and controls in which untransformed cells were used
as a target. The PCR were performed as described above. except that the pnmcr
annealing temperature was higher (68°C). To screen for SSU rDNA diversity. the
ampliﬁed inserts were ﬁrst digested ovcrntght at 37°C by adding 0.2 U of HhaI
and 0.2 U of Haelll (Gibco BRL. Gaithersburg. Md.) to 5 ul of the PCR product.
The resulting fragments were separated by gel electrophoresis in 3.5% Metaphor
agarosc (FMC Bioproducts. Rockland, Maine) in the presence of cthidium
bromide and fresh 1x TBE buﬁcr (22) at 4'C and 5 Wm for 6 h. Clones with
identical restriction patterns were digested with two additional tetramcric re-
striction endonucleases (0.2 U of Mspl and 0.2 U of Rsal). The similarities
between the elcctrophorctic patterns of restriction fragments were analyzed with
GelCompar software (Applied Maths. Kortrijk, Belgium). The cluster analysis
method used was comparative numerical analysis with the unweighted pair group
method using arithmetic averages (UPGMA). Individual clones were grouped by
using a cutoff of 97% similarity and a 5% error rate for the band position. The
diversities of the phylotypes in different samples were compared by rarefaction
analysis (13, 30).

DGGE. Denaturing gradient gel electrophoresis (DGGE) was performed as
previously described by Muyzcr (24) with eubacterial PCR primers F-968 and
R-l401 of Nubcl et al. (26). The parallel denaturing gradient was cast with
denaturing agent concentrations ranging from 0 to 60%. The fragments were
made visible by acidic silver staining.

Determination of nucleotide sequences and phylogenetic analysis. Ampliﬁed
SSU rDNA clone inserts were puriﬁed (Wizard PCR Preps: Promega) and
partially sequenced. Nucleotide sequences were determined by the fluorescent
DiDeoxy termination method by performing automated ﬂuorescent Taq qcle
sequencing with the ABI Catalyst SOC-AB! 373A sequencing system (Applied
Biosystcms, Foster City. Calif). To ensure accuracy and to aid in chimera
detection. borh ends of the SSU rDNA molecule were sequenced with reverse
pnmers 1529171 (S’-CGCGGCI‘GCTGGCAC-3') and rP2 (41). All sequences
were about 400 bases long and were aligned manually with sequences in the SU
database of the Ribosomal Database Project (RDP) (21) based on primary- and
secondary-structure considerations The results obtained were compared to
alignments obtained with Align Sequence. version 2.0, from the ARB sequence
analysis software package (33). Phylogenetic relationships were inferred by using
the neighbor-joining method (28) and the modiﬁed Juices-Cantor algorithm (16.
42). The robustness of the ﬁnal rcpology was tested with the tree-building
methods PAUP (34) and fastDNAml (21). All phylogenetic assignments were
made and phylogenetic trees were constructed within the ARB software package
(33) by using version 5.0 of the RDP database (21). Only unambiguously aligned
regions were used for the sequence analysis (Table 1).

To detect potential chimeric artifacrs in the partial sequences of the 3' end and
the 5' end. as well as the entire SSU rDNA gene. two strategies were used. The
partial sequences that were around 400 bases long were (i) examined with the
CHECK_CHIMERA program oﬂcred by RDP (21) and. for comparison. (ii)
examined with the mglobalCl-il program offered by the USC Computational
Biology web site (17). To detect potential chimeric artifacts in a complete SSU
rDNA gene. the phylogenies determined from the sequences of the 3' ends and
the 5' ends were compared.

Nucleotide sequence accession numbers. The sequence data for Hawaiian
rainforest soil clones HRS-1 through HRS-23 have been deposited in the Gen-
Bank database under accession no. AF016514 to AF016533.

RESULTS

Soil analysis. The chemical analysis of the soil revealed a
composition quite typical for a very young soil compared to
similar but older forest soils. The soil organic matter content
was 3.1%, and the pH was 5.4. The soil nitrate N and ammonia
N contents were determined to be 0.2 and 3.6 ug/g, respec-
tively, while the total nitrogen content was 0.08%. The sand
particles had sharp faces, which indicated that there had been
little weathering and resulted in rapid water inﬁltration. The
humic material had to be less than 200 years old, and hence its
chemical structure was different than that of typical soil humic
acid.

In this young tr0pical rainforest soil the cm ratio (a mea-
sure of biological activity in soils) was 16:1 (compared with a
global average of 14:1), suggesting that an active bacterial
community was present. This assumption was supported by
high leaf litter decomposition rates (37), as well as elevated in
situ mineralization and nitriﬁcation rates (6). The cation ex-
change capacity was low (10.8 meq/lOO g) due to primary
minerals.

29

VOL 64. 1998

rRNA- CHARACTERIZATION OF HAWAIIAN SOIL 1285

TABLE I. Phylogenetic afﬁliations based on SSU rDNA genes of members of a Hawaiian rainforest soil bacterial community'

 

 

 

 

Relative Most closely related organism in RDP database ‘7” Similarity ‘0
Phylotype abundance Phylogenetic aﬁiliation mostlcloscly
(%)° Taxon RDP subdivision 0:: “F
gamsm
35% G+C fraction
HRS-l 6.8 Clostridia and their relatives [Closm'dium Ierani] Clostridium novyi subgroup 80.6
HRS-2 6.8 Clostridia and their relatives [Clostridium Ietani] Closmdium novyt' subgroup 76.1
HRS-3 6.8 Clostridia and their relatives Closmdium fallax Closmdium perfringens assemblage 92.9
HRS-4 6.8 Clostridia and their relatives Clostﬁdiwn fallax Clostridia”: perfn'ngens assemblage 94.0
HRS-6 5 l Clostridia and their relatives Clostridium puniceum Closmdium buryricum subgroup 95.3
HRS-7 l 7 Clostridia and their relatives Clostridium buryn'cum Closm'dium butyn'cum subgroup 94.7
HRS-8 1.7 Clostridia and their relatives Closrn'dium puniceum Clostridium bulyricum subgroup 98.0
HRS-10 1.7 Clostridia and their relatives Clostridium beijerinckii Clostridr‘um buryn'cum subgroup 98.5
HRS—21 1.7 Clostridia and their relatives Clostridium puniceum Clostridium buryn'cum subgroup 98.5
HRS-22 1.7 Clostridia and their relatives [Clostridium tetanomorphum] Closnidium buryricum subgroup 80.0
63% G+C fraction
HRS—12 13.2 Alpha purple bacteria [Rhodopseudomonas viridis] Rhizobium-Agmbacreriwn group 82.6
HRS-13 10.5 Gamma purple bacteria Pseudomonassyrm syrrngae Pseudomonas subgroup 97.9
HRS-16 7.9 Alpha purple bacteria Azospinllum lipoferum Rhodospirillum mbrum assemblage 88.0
HRS-17 1.3 Alpha purple bacteria Zoogloea mm' Brucella assemblage 90.4
HRS-18 1.3 Fibrobacier phylum Environmental strain MC Acidobacten‘um subdivision 93.2
03
HRS-19 1.3 BacillrLr-Lactobacillur [Bacillus ﬂavorhermus] Bacillus megarerium group 80.3
subdivision
HRS-20 1.3 Delta purple bacteria Environmental strain FIE 20 Myxobacterium group W9
'Only unambiguously aligned regions were used in the anaislys

’ Relative abundance of clones belonging to a phylotype. calculated by dividing the number of clones belonging to the phylotype by the total number of clones

III-lined
‘ Buckets indicate that the taxonomic asignment of the closest relative is uncertain (level of similarity, less than 85%).

DNA extraction and puriﬁcation. Direct microscopic cell
counts of soil smears before and after DNA extraction re-
vealed a high lysis efﬁciency, 91% I 3%. The lysis efﬁciency as
estimated by DNA yield was also high; 6.4 pg of DNA/g (dry
weight) of soil was recovered. compared to an expected DNA
yield of 5. 8 ug/g, which was calculated by multiplying9 the mi-
croscopic bacterial cell counts (1.6 x 10 I 0.3 X 10 g9cells/g)
by an average of 4 X 10'” g of DNA cell‘1 (9). Coextraction
of humic material prevented PCR ampliﬁcation of SSU rRNA
genes if the standard puriﬁcation protocol of Holben was used,
probably because of the unusual nature of the young humic
acids. Additional clean-up by preparative agarose gel electro-
phoresis and an additional series of washing steps in a centrif-
ugal concentrator were needed to obtain DNA clean enough
for reliable ampliﬁcation of the bacterial SSU rRNA genes.

Community G+C proﬁle. A proﬁle of the community com-
position was obtained based on the amounts of DNAs having
different G+C contents. The G+C contents of the majority of
the soil DNA were in the range from 52 to 68% (Fig. 1); this
included members of genera known to dominate soil bacterial
communities, including the genera Agrobacren‘um (57 to 63%
G+C),A1caligenes (56 to 63% G+C),Anhmbacrer (63 to 69%
G+C), and Pseudomonas (58 to 66% G+C) (12). A rather
consistent but minor quantity of DNA was found with G+C
contents ranging from 30 to 50%, a range which is found in
members of soil genera like Streptococcus (35 to 40% G+C),
Clot-iridium (24 to 54% G+C), and Bacillus (32 to 69% G+C).
Compared with the community proﬁles of midwestem agricul-
tural soils, the main peak of the Hawaiian bacterial community
proﬁle was shifted toward a lower G+C content (by approxi—
mately 4% G+C) (Fig. 1).

Phylotype abundance patterns. Agarose gel electrophoresis
revealed a clean band of SSU rRNA genes selectively ampli-
ﬁed from puriﬁed soil-extracted DNA. The clones were di-

gested with restriction enzymes (ampliﬁed rDNA restriction
analysis) (Fig. 2) and sorted by cluster analysis. Of the 81
clones obtained from unfractionated soil DNA that tested pos-
itive for alpha-complementation of B-galactosidase, 67 con-
tained the 1.5-kb SSU rDNA insert. Primary restriction with
Hull] and Hhal resulted in 64 diﬁerent restriction patterns
(Fig. 2 and 3A), indicating that there was a high level of
diversity. Clones with similar restriction patterns were not dif-
ferentiated when the preparations were further digested with a
combination of Mspl and Karl. The two patterns that were
repeated each accounted for only 3% of all of the SSU rDNA
clones, while the remaining 60 patterns were each represented

5 of total DNA

y
53
d'
a!
s
u
30 40 SO 60

 

$6“:

5- I a

“6.1.” 'dptprminr-d
by the G+C content of its DNA. The bacterial community proﬁle of a mid-
Miehigan agricultural soil is included for comparison (12 )

1286 NUSSLEIN AND TIEDJE

Kidd???

sffffsff

30

APPL. ENVIRON. MlGlOBlOL.

#45593? _I"I"I‘f,"‘.i

siffaffffff

 

2. Restriction patterns of ampliﬁed SSU rRNA clones in the 63% G+C fraction after restriction digestion with Ham and ”MI. Plasmid pBR322 digested

with Haelll (marker V) was used as a DNA size marker.

by a single clone (Fig. 3A). In order to reduce this diversity to
a more manageable level, two G+C content fractions (35 and
63% G+C fractions) of the whole community DNA were used
in a similar analysis. We chose the 63% G+C fraction because
it was located within the major peak of DNA abundance typ-
ically found in temperate region soils. The second fraction, the
35% G+C fraction, was chosen randomly to represent a

tion of the minor members of the community (Fig. 1). Table 1
shows the relative abundance of selected clones as a measure
of dominance relative to the entire clone library examined. The
63% G+C fraction produced 46 diﬂerent patterns for the 76

Clone“) per Phylotype

3° ‘0
Phylolypes

. 3. Frequenq' distribution of SSU rDNA gene phylotypes (restriction
patterns) from the total soil DNA (A). the 35% G+C fraction (B), and the 63%
G+C fraction (C) of a young Hawaiian soil. The proﬁles are based on results
obtained after digestion with tetrameric restriction endonucleases Hulll plus
Him] and Mrpl plus Rsal.

 

clones examined (Fig. 3C). Eleven of these patterns were rep-
resented by two or more clones, which together accounted for
54% of the clones investigated. The three most dominant re-
striction patterns, patterns 1, 2, and 3, accounted for 13, 11,
and 8%, respectively of the SSU rDNA inserts analyzed. The
35% G+C fraction produced 47 different patterns for the 59
clones examined (Fig. 313). Only seven of these patterns were
represented by two or more clones, and they accounted for
32% of the clones examined. The two most dominant patterns
each represented 7% of the SSU rDNA insert diversity in this
fraction. In both fractions together only two clones with similar
resn'iction patterns were differentiated when the preparations
were digested with the second restriction endonuclease pair,
MspI plus RsaI.

The quantitative effect of using a fraction of the community
DNA instead of the total community DNA on the diversity of
phylotypes was evaluated by rarefaction analysis. Given the
species abundance distribution of a clone library, rarefaction
gives estimates of the species richness of subsamples taken
from it. This analysis veriﬁed that the unfractionated DNA
contained too many phylotypes to reveal any structure but that
fractionation of the DNA on the basis of G+C content did
reduce the diversity to a level at which structures of dominance
could be detected; the data also suggested that the phylotype
sampling in the two fractions was far from complete (Fig. 4).

DGGE analysis. The abundance of particular SSU rDNA
clones in our clone library may not represent the actual quan-
titative abundance of the clones in the soil sample due to a
PCR or cloning bias. To assess this, we used DGGE to deter-
mine whether the dominant phylotypes in the clone library
corresponded to well-represented phylotypes on DGGE gels
of community DNA. Parallel analysis by DGGE of the two
G+C fractions and dominant clones from the two fractions
revealed that the intensely stained bands in the community
DNA in each G+C fraction (indicating strong representation)
corresponded to the bands obtained from the individual clones

V01. 64, 1998

 

    
   

—.— A: Total Soil DNA

>. 60f -l 3:35‘iGoCFmtion
E

2 so-
G

S.

2‘ 40'
e

3'.

= 1
ﬁ- 30.
h-

e

3

2 20
E

3i

" rol‘

 

0 IO

20 30 40 50 60 70
Number of Clones
FIG. 4. Rarefaction urn/es for the phylotypes found in total soil DNA and in
the 35 and 63% G+C fractions of a Hawaiian soil. The expected numbers of

phylotypes calculated from a random sample of individuals taken from the total
population of phylotypes are shown on the y axis.

(Fig. 5). Even though phylogenetically unrelated strains can
have bands with the same R, value due to identical melting
behavior of the SSU rDNA fragment, it is unlikely that a strain
other than the one detected was both strongly represented and
had the same R, value.

Sequence analysis. Two representative clones of the ﬁve
dominant phylotypes from both abundance proﬁles (Fig. 3)
and selected rare individuals were used for a partial sequenc-
ing and phylogenetic characterization analysis. This analysis
revealed corresponding phylogenetic afﬁliations for clones be-
longing to the same phylo (e.g., clones HRS-1 and HRS-2
and clones HRS-3 and HRS-4) (Table 1), although the clones
varied somewhat in sequence similarity (not all matching pairs
are shown in Table 1) Of the 23 clones sequenced 20 were
members of the domain Bacteria, while three clones were dis-
missed as possible chimeras. The phylogenetic aﬂiliations and
closest relatives in the RDP database are shown in Table 1.
Figure 6 illustrates the phylogenetic relationships among some
of the 35% G+C fraction clones. Several of the clostridial
clones are closely related to each other. The phylogenetic
aﬂiliations of particular clones and their levels of abundance
are summarized in Fig. 7. None of the clones exhibited an exact
match with any of the SSU rDNA sequences found in the
databases. In particular, clone HRS-18, which is related to the
Acidobacraium subdivision, conﬁrmed that novel taxa discov-
ered previously in other molecular surveys of soil were present
(20).

DISCUSSION

Geographic isolation, young age (200 years), constant cli-
mate, and low diversity of plant species did not reduce the
microbial diversity in the soil studied to an extent that revealed
more than two instances of resampling of the same eubacterial
phylotype in a 70-clone library. Hence, the diversity was too
great to reveal underlying community structure by this method.
Similarly, Bomeman et al. (4) found only 4% duplicates among
124 soil rDNA clones from an older continental soil. One
approach to reduce complexity is to limit the study to only
certain subsets of the community, an approach typically used
by macroecologists. We attempted to do this by analyzing
rDNA fractions having certain G+C contents since particular
bacterial taxa have characteristic G+C contents. Clone librarv

31

rRNA CHARACTERIZATION OF HAWAIIAN SOIL 1287

12345

l-ur

k“. .

.. -.- non?"

I
r
l

 

FIG. 5. DGGE analysis of SSU rDNA fragments (length. 434 bp) obtained
after PCR ampliﬁcation of the 35 and 63% 6+ C fractions of soil DNA and
individual SSU rDNA clones from the clone libraries that were most frequent as
determined by ampliﬁed rDNA restriction nanalysrs islane 1 clone HRS-2 of the
35% G+C fraction; lane 2. 35% G+C fraction; lane 4 63% G+C fraction: lane
5. clone HRS- 12 ofthe 63% 0+" franinnl am
bacterial genomic DNAs as a marker and positive control. The ﬁgure is a
negative image of a silver- -staine DGGE separation pattern. The arrows indicate

 

tive soil DNA fractions. PCR products obtained from some strains (lane 5)
produced more than one band due to sequence heterogeneities of 165 rRNA
operons (24).

ies obtained from pooled DNA fractions (e.g., DNAs having
G+C contents of 61 to 65%) still contained few repeated
phylotypes (data not shown), but samples from one of these
fractions (e.g., the 63% G+C fraction) contained more re-
peated phylotypes, indicating that more complete coverage of
the rDNA types in this sample was obtained. Rarefaction anal-
ysis also indicated that there was reduced diversity in the in-
dividual G+C fractions compared to the total DNA, especially
the 63% G+C fraction. but that rDNA diversity wu far from
exhausted in a 76-clone library of this fraction (Fig. 4). Hence,
separation on the basis of G+C content revealed new diversity
and provided evidence that soil rDNA diversity is much greater
than the diversity that is revealed by eubacterial clone libraries
of total community DNA. This estimate of greater soil rDNA
diversity supports the high level of bacterial diversity estimated
by Torsvik et al. (36) based on rates of soil DNA reannealing
The G+C content separation method also offers a way to
enrich for rarer members of the community since DNAs having
other G+C contents. especially DNAs of dominant types, can
be removed by separating the DNAs into dilIerent fractions.
Dominance of phylotypes was observed in the 63% G+C
fraction but not in the 35% fraction. This difference is consis-
tent with the ecological prediction that the most dominant

1288 NUSSLEIN AND TIEDJE

dnorlqns
mnaIMurq
WMPI‘IIOIJ

53$
.,.. iii
:gg 9
is: .i
F:

 

 

FIG. 6. Phylogenetic relationships of the most closely related SSU rDNA
clones from the 35 % G+ C fraction of the young Hawaiian soil used. Evolutioln-
ary dis nces were determined by maximum- -liltelihood analysis. Bar:
substitution per base position. env., environmental clone.

biomass (i.e., the 58 to 65% G+C fraction) reflects the most
competitive organisms, which consist of fewer species (19). The
lack of dominant phylotypes in the less successful fractions
(e.g., the 35% G+C fraction) is consistent with the expectation
that the diversity in the secondary populations is greater (14).

The nucleotide sequence analysis of the rDNA clones iden-
tiﬁed taxa that were expected for DNAs having G+C contents
of 63 and 35%. Since G+C content is more conserved in the
mt operon than in the genome as a whole (39), the separation
method must be driven primarily by the G+C content of the
ﬂanking DNA. The DNA fragments obtained by the DNA
isolation method used were usually more than 20 kb long.

The aerobic bacteria most often cultured from soil (e.g.,
members of the genera Anhrobacter, Pseudomonas, and Burk-
holden‘a) have the same G+C contents as the most dominant
biomass, as determined from the DNA data. Since only a low
percentage of soil bacteria can be cultivated, these data suggest
that the majority of the unculturable types must have DNA
G+C contents of 55 to 68 .o, and hence these organisms are
either close relatives of the typical culturable forms or are new
types buth ave G+C contents in thisran rane.g

Finding a large diversity of clostridia' in the 35% G+C frac-
tion was initially unexpected since the site used rs well drained
with high aeration and porosity (Fig. 6). However, the high
rainfall throughout the year and the young organic matter
could provide anaerobic microsites for growth of the clostridial
community (1), and bird feces is one of the most feasrble
inoculum sources for this group.

The microbial colonization of the young Hawaiian soil ex-
amined by very diverse phylotypes rs not easily explained es-
pecially considering the very high rDNA diversity in the G+C
fractions. The geographic isolation of the Hawaiian Islands
certainly limited colonization by plants (seed dispersal), in-
sects, and other organisms, including humans (until 500 AD).
Bacterial colonization of soil could have resulted from avian
transport, including avian fecal introductions, from seawater
spray, or from aerial transport. Aerial transport has been con-
sidered very inefﬁcient due to poor microbial survival caused
by long exposure to UV light, desiccation, and the likely fallout
of any microbe-associated soil particles during transport from
Asia (more than 10,000 km). Recently, however, sand grains
from sandstorms in the Gobi Desert and loess plateau regions

Arm. ENVIRON. Mrcrtonror.

_Fraetionz 35% GoC
It.

Gas and Fblalrves/Gos novvr Sub mu

m—
'Ocs hulvr Gou 'Clos. 09mm ens Assemble

 

      

 

 
  
  
    
  
  

bu VYCUY mSu
r—u—
Gas and Peiatwesl
’Oos Durw. Group

rDNA Clones [Pattern

   

 

 

 

 

 

 

 

 

E r Alpha-Pur-Ial Fhlzobmm-A-Iobacl. Gouo

E Mma-Purpld F's-adorn. . a rou-

\

- 1 Al- M-Puethodo. rubrurn Assamb -

é Alpha-Purple] ﬁbmboctor Blenucrnnac. Swdv.
Bruce/Ia Phylum Bic. mogumium

g Assemblage Acidobact. stream Group

a

.

 
 

$1

1 10 20 30 40
Mrletlon Pattern (Phylotype)

FIG. 7. Phylogenetic affiliation of the dominant and rare phylotypes identi-
fied in phylotype abundance distribution proﬁles for the 35 and 63% G+C
fractions of Hawaiian soil. The numbers indicate the designations of the clones
that were analyzed (1, HRS—l; 2,HRS- -2.etc). Abbreviations. Clos., Closm‘dt'um;

los. butyr.,Closaidiumbutyr-rcum; Agrvbcct. ,Agrobacwium; Pseudom, Pseudo-
Mr' ;Acr'dobact. Subdiv., Aadobacmiumy subdivi~
sion; wBac- Lacrabac. Subdiv., Bacillus- Lactobaa'llm subdivision, Bac., Bacillus.

of central People’s Republic of China were tracked to Hawaii,
even though the density and the size of the sand grains sug-
gested that they should have been deposited long before they
reached Hawaii (2, 15,23). Regardless of the sources of bac-
terial colonization of Hawaiian soil, the soil communities ap-
pear to be unexpectedly complex, but this does not mean that
they display the full complement of microbial types or diversity
found in montane rainforest soils of continental environments.

Studies such as this one, performed by using PCR ampliﬁ-
cation of SSU rDNA genes from community-derived DNA.
should never be assumed to be comprehensive because of
well-known biases (17, 39). In particular, they are likely to
reﬂect rm operons that are more readily PCR ampliﬁable.
Nevertheless, this study supports the notion that even young
terrestrial environments exhibit enormous diversity and con-
tain novel, uncultivated organisms.

ACKNOWLEDGMENTS

We thank T. Hattori and J. Urbance for comments on the manu-
script and P. Lepp for instructions on use of the ARB software.
th Harris for his introduction to the G+C fractionation tech-
niquc and microscopic image analysis.

This work was supportc ed by National Science Foundation grant
DEB 9120006 to the Center for Microbial Ecolo

REFEENCES

1. Arab. J. R. M. and K. A. Smith. 1985. Anaerobic micro-environments in soil
and the occurrence of anaerobic bacteria, p. 247-261. In V. Jensen. A.
Kjoller and 1.. H. Sorensen (ed. ), Microbial communities in soil microbiol-
ogy. FJsevier Applied Science Publishers. London, Englan
.J’Betaer .R..K. LCarder,R. Duct, .l.T .Merill, N. W. ‘l‘intlale. M.
”Unnatsu. D. K. Costello. R. W. Young, R.A. A.Feely. .l A. Breland,R.
Bernstein. and A. M. Green. 1988. Long-range transport of giant mineral

VOL. 64. 1998

11.

aerosol particles Nature 336568-571.

. Boivin-Jahns, V.. A. Bianchi. R. Rnimy, J. Garcia. S. Daumas. and R.

Christen. 1995. Comparison of phenotypical and molecular methods for the
identiﬁcation of bacterial strains isolated from a deep subsurface environ-
ment. Appl. Environ. Microbiol. 61:3400—3406.

. Bomeman. J» P. W. Sltroch. K. M. O'Sullivan. J. A. Panlus. N. G. Rum-

janek. J. L. Jansen, J. Nienhns. and E. W. Triplett. 1996. Molecular micro-
bial diversity of an agricultural soil in Wisconsin. Appl. Environ. Microbiol.
62:1935—1943.

. Carlquist, S. C. 1980. Hawaii. a natural history. Paciﬁc Tropical Botanical

Garden. Honolulu. Hawaii.

. Crews, T. E. K. Kitayama. J. H. Fmes. R. H. Riley. D. A. Herbert, D.

Mneller-Dombois. and P. M. Vitousek. 1995. Changes in soil phosphorous
fractions and ecosystem dynamics across a long chronosequence in Hawaii.
Ecology 76:1407-1424.

. D’Aquila. R. 1., L. J. Bechtel, J. A. Videler. J. J. Eron. P. Goraca. and J. C.

Kaplan. 1991. Maximizing sensitivity and speciﬁcity of PCR by preampliﬁ-
cation heating. Nucleic Acids Res 19:3749.

Department of Geography. University of Hawaii. 1983. Atlas of Hawaii. 2nd
ed Universrty of Hawaii Press. Honolulu.

Ellenbeoek. F. M. and T. E. Cappenberg. 1991. DNA synthesis and tritiated
thymidine incorporation by heterotrophic freshwater bacteria in continuous
culture. Appl. Environ. Microbiol. 57:1675—1682.

. Gianabelluca, T. W. M. A. Nullet. and T. A. Schroeder. 1986. Rainfall atlas

of Hawaii. Department of Land and Natural Resources. Honolulu. Hawaii.
Holben. W. E. 1994. Isolation and puriﬁcation of bacterial DNA from soil. p.
727-751. In R. W. Weaver et al. (ed). Methods of soil analysis. part 2.
Microbiological and biochemical properties. Soil Science Society of America
Book Series no. 5. Soil Science Society of America. Madison. Wis.

12. Holhen. W. E. and D. Harris. 1995. DNA-based monitoring of total bacterial

13.
I4.

15.

community structure in environmental samples Mol. Ecol. 4:627-631.
Hurlbert. S. H. 1971. The nonconcept of species diversity: a critique and
alternative parameters. Ecology 52:577-586.

Huston. M. A. 1994. Biological diversity. Cambridge University Press, Cam-
bridge. United Kingdom.
Jaekson.M.L,T.W.M.Levelt.J.K.Syers.R.W.Rex.R.N.Clayton,G.D.
Sherman. and G. Uehara. 1971. Geomorphological relationships of tropo-
spherically derived quartz in the soils of the Hawaiian Islands. Soil Sci. Soc.
Am. Proc. 35:515-525.

. Jokes. ‘1'. H. and C. R. Cantor. 1969. Evolution of protein molecules. p.

21-132. In H. N. Munro (ed). Mammalian protein metabolism. Academic
Press. New York. NY.

. Kauatsonlis, G. A., and M. S. Waterman. 1997. A new computational

method for detection of chimeric 16$ rRNA artifacts generated by PCR
ampliﬁcation from mixed bacterial populations. Appl. Environ. Microbiol.
639338-2346.

. [altar-ea, C.. and K. Paigen. 1980. A simple. rapid and sensitive DNA assay

procedure. Anal. Biochem. [02:344-352.

. Levine, S. H. 1976. Competitive interactions in ecosystems. Am. Nat 110:

903-910.

. Bunch. W. and E. Stackebrandt 1992. Occurrence of novel groups of the

domain Bacteria as revealed by analysis of genetic material isolated from an
Australian terrestrial environment. J. Bacteriol. 174:5072-5078.

.Maldolr,B.L.G.J.Olsen.N.hrsekawMM.J.MeCaoghey.and

C. R. Woese. 1996. The Ribosomal Database Project. Nucleic Acids Res
14:82-85.

22. Maniatis. ‘I'.. E. F. Fritsch, and J. Sanbroolc 1982. Molecular cloning: a

23.

24.

laboratory manual. Cold Spring Harbor Laboratory. Cold Spring Harbor.
N.Y.

Merrill, J. T. 1989. Atmospheric long-range transport to the Paciﬁc Ocean.
p. 15—50. In J. P. Riley and R. Chester (ed.), Chemical oceanography. vol. 10.
Academic Press. New York. NY.

Mnyaer. G. 1995. Protocols for the DGGE. p. 8—14. In K Smalla and G.
Muyzer (ed.). EU-Workshop on the Application of DGGE and TGGE in
Microbial Ecology. Braunschweig. Germany. Biologische Bundesanstalt fur

33

25.

26.

2'7.

29.

31.

rRNA CHARACTERIZATION OF HAWAIIAN SOIL 1289

Land- und Forstwirtschaft, Braunschweig. Germany.

Nannipieri. P. S. Green, and B. Ceccanti. 1990. Ecological signiﬁcance of the
biological activity in soil, p. 293-356. In J.-M. Bollag and G. Stotzky (ed.).
Soil biochemistry. vol. 6. Marcel Dekker, Inc., New York. NY.

Niibel. U., B. Engelen. A. Felske. J. Snaidr. A. Wieshuber. R. L Amann, W.
laidwig, and H. Backhaus. 1996. Sequence heterogeneities of genes encod-
ing 16$ rRNAs in Paem‘bacz’llus polvmvxa detected by temperature gradient
gel electrophoresis. J. Bacteriol. 178:5636—5643.

Peck. T. R.. D. Megel. K. Eik. D. A. Whitney, D. Warncke, R. C. Minter, J. R.
Brown. D. Knudsen. W. C. Dahnke. D. J. Eckert, D. Beegle. P. E. Finn, and
E. E. Schulte. 1988. Recommended chemical soil test procedure for the
north central region. North Dakota Agricultural Experiment Station. North
Dakota State University, Fargo.

. Saiton. N. and M. Nei. 1987. The neighbor-joining method: a new method

for reconstructing phylogenetic trees MOI. Biol. Evol. 4:406—425.
Schmidt, T. M.. B. Pace. and N. R. Pace. 1991. Detection of DNA contam-
ination in Taq polymerase. BioTechniques 11:176-177.

. Sirnberlol, D. 1978. Use of rarefaction and related methods in ecology. p.

ISO-165. In K. L Dickson. .1. Cairns, Jr.. and R. J. Livingston (ed). Biolog-
ical data in water pollution assessment: quantitative and statistical analyses.
American Society for Testing and Materials. West Conshohocken. Pa.

Soil Survey Stat. 1972. Soil survey of the islands of Kauai. Oahu, Maui,
Molokai. and Lanai. state of Hawaii. US. Government Printing Ofﬁce,
Washington, DC.

32. Sarheim. R.. V. L Torsvik, and J. Golrsoyr. 1989. Phenorypical divergences

33.

35.

37.

39.

40.

41.

between populations of soil bacteria isolated on different media. Microb.
Ecol. 17:181-192.
Strunk.0.,O.Gross,B.ReicheI.M.May,S.Her-ann.N.Stncknaan.B.
Nonhol.M.Lenke.A.Ginhart,T.Lndwig.A.Bode.K.—H.Schleiter.aadw.
Ludwig. Arb: a software environment for sequence data. Nucleic Acids Res,
in press.

. Swolord, D. L 1990. PAUP: phylogenetic analysis using parsimony. version

3.0. Illinois Natural History Survey. Champaign.
Tiedij.M..J.-Z.Zhon,K.Nissletn.C.LMoyer.andR.R.Faltborpe.
1997. Extent and patterns of soil microbial diversity. p. 35-41. In M. T.
Martins et al. (ed), Progress in microbial ecology. Brazilian Society for
Microbiology. Sao Paulo. Brazil.

. Torsvik. v. J. Gokseyr. and r. 1. Dane. 1990. High diversity in DNA of soil

bacteria. Appl. Environ. Microbiol. 56:782-787.

Vitousek. P. M. D. R. Turner, and K. Kit-yam. 1995. Foliar nutrients
during long-term soil development in Hawaiian montane rain forest. Ecology
76:712—720.

.Wand.D.M..M.M.Bateson,R.Weller.aadA.LRnl-Robens. 1990. 165

rRNA sequences reveal numerous microorganisms in a natural community.
Nature 345:63—65.
Ward,D.M.MM.Bateaon,R.WeIler.andA.LRal-Roberts.l992.
Ribosomal RNA analysis of microorganisms as they occur in nature, p.
219—286. In K. C. Marshall (ed.). Advances in microbial ecology. vol. 12.
Plenum Press. New York. NY.

Weisblnm. B. and E. Haenssler. 1974. Huorornetric properties of the bis-
benzimidazole derivative Hoechst 332.58. a fluorescent probe speciﬁc for AT
concentration in chromosomal DNA. Chromosoma 46:255-260.
Weisbnmw.G.S.M.Barns.D.A.Pelletier.andD.J.Lane.I991.16$
ribosomal DNA ampliﬁcation for phylogenetic study. J. Bacteriol. 173:697-
703.

41 Weisbnrpw.G..J.G.Tully,D.LRooe,J.P.PetaeI,H.Oyaiu.D.Yang.L

43.

Mandelco.J.Sechrest.T.G.I.awrenee.J.VanEnen.J.Manilol.aadC.R.
Woese. 1989. A phylogenetic analysis of the mycoplasrnas: basis for their
classiﬁcation. J. Bacteriol. 171:6455-6467.

hon. J., M. A. Bruns, and J. M. 'I‘iedje. 1996. Rapid method for the recovery
of DNA from soils of diverse composition. Appl. Environ. Microbiol. 62:
316—322.

.Zhon.J.M.E.Davey,J.B.ﬁgnens.E.Rivkina.D.Glllchinsky.andJ.M

Tiedje. 1997. Phylogenetic diversity of a bacterial community determined
from Siberian tundra soil DNA. Microbiology “35913-3919.

CHAPTER 3

SOIL BACTERIAL COMMUNITY SHIFT LINKED TO CHANGE
FROM FOREST TO PASTURE VEGETATION IN A TROPICAL SOIL

ABSTRACT

The change in vegetative cover of a Hawaiian soil from forest to pasture led to
signiﬁcant changes in the composition of the soil bacterial community. DNA was
extracted from both soils and compared by the abundance of its guanine and cytosine
(G+C) composition, by analysis of abundance of phylotypes of small-subunit
ribosomal DNA (SSU rDNA) ampliﬁed from 63% and 35% G+C fractions, and by
phylogenetic analysis of the dominant rDNA clones in the 63% G+C ﬁaction. All three
methods showed differences between sites that were greater than within sites, providing
evidence that vegetation had a strong inﬂuence on microbial community composition at
several levels of taxon resolution. The forest soil DNA had a peak in G+C content of
61% while the DNA of the pasture soil was shifted to higher G+C content with a peak
at 67%. The phylotype abundance profiles of the 63% G+C fractions showed more
dominance in the forest soil and greater phylotype richness at the pasture soil. None of

the dominant phylotypes found in the forest soil could be detected in the pasture soil.

34

35

In the 63% G+C fraction sequence analysis of the three most dominant members
revealed that their taxa changed from F ibrobacter and Syntrophomonas assemblages in
the forest soil to Burkholderia and Rhizobium-Agrobacterium assemblages in the

pasture soil.

INTRODUCTION

Dispersal of inoculum and selective forces for growth provided by the
environment are the two main factors which determine which organisms dominate a
habitat. Vegetation is one environmental factor thought to be a major determinant of the
composition of the soil microbial community since it provides the primary resource for
heterotrophic growth. Since different plant species are comprised of different carbon
compounds, different microorganisms might be expected to grow in different plant
communities. Furthermore, the plant may alter other physical and chemical features of
the soil and hence favor the growth of different species. Little data exists, however, on
the inﬂuence of vegetation on the composition of soil microbial communities.

We investigated the inﬂuence of vegetative cover on the structure and
composition of a microbial community developed in a volcanic ash soil on the Big Island
of Hawaii. Because of its young age and geographic isolation, Hawaii is depapurate in
macrobiological species providing a simpler ecosystem for study. Two adjacent soils

from the Kohala region were compared, one which has been continuously covered by a

36

native tropical forest, and the other covered by a grass pasture, which replaced the rain
forest approximately 80 years ago. DNA based methods were used rather than culturing
methods to compare the soil communities since they better recover the dominant
prokaryotic populations. Furthermore, community DNA allows analysis at different
levels of taxon resolution and provides more quantitative comparisons than culture
based methods. The three levels of resolution used in order of increasing speciﬁcity and
decreasing comprehensiveness are as follows. First, community DNA was fractionated
by its guanine and cytosine (G+C) content and the amount of DNA of each G+C
content quantiﬁed. Second, DNA fractions of two different G+C contents, one from a
fraction of high biomass (63% G+C) and the other of low biomass (35% G+C) were
then selected for analysis of small subunit ribosomal RNA (SSUrRNA) genes ampliﬁed
by PCR. The rDNA clone libraries were screened by ampliﬁed ribosomal DNA
restriction analysis (ARDRA) to determine pattern abundance proﬁles. Third, the
dominant clones in the 63% biomass fraction from each soil were fully sequenced and
analyzed phylogenetically. Together these three methods provide information at
different levels of resolution on similarities and differences in the microbial communities
from the two soils. Differences in the bacterial community due to vegetation were seen

by all three methods.

37

MATERIALS & METHODS

Soil origin and sampling. The soil is located within the Kohala Forest
Reserve on the Big Island of Hawaii (lat/long. 20°03’/155°41’) and has never been
cleared by humans. The forest site is the same as the Kohala site in Vitousek’s
chronosequence (Chapters 2 and 4). The pasture site was cleared in the 1920’s for
ranchland, and lies immediately adjacent to the Kohala rainforest. The distance between
sampling sites was approximately 100m. The pasture was seeded with the Aﬁ'ican
grass, Pennisetum clandestinum, which now dominates this species-poor site. Cattle
graze the pasture.

The parent material of this Typic Placandept is a volcanic ash deposit, tephra,
and is 150,000 years old. The sampling sites are at 1122 m elevation which corresponds
to a mean annual air temperature of 16°C (Atlas of Hawaii, 1983). The mean annual
rainfall is near 2500 mm (Giambelluca et al., 1986). The pristine closed canope
rainforest is mono- dominated (83% cover) by the native tree Metrosideros polymorpha,
the most widespread tree on the Hawaiian islands (Kitayama et al., 1996). The well
developed understory vegetation is dominated by native tree ferns Cibotium spp. (25%
cover). Other genera of trees and shrubs at this site are Cheirodendron (18% cover), the
understory tree Coprosoma (4% cover), and the shrub Vaccinium (2% cover), account
for most of the remaining cover (Crews et al., 1995). The exotic plant species Psidium
cattleianum and Hedychium garnerianum were also present in low abundance (Riley and

Vitousek, 1995). The total number of vascular plant species per 0.2 ha is 53, while the

 

 

38

native plant species diversity is 45 (Crews et al., 1995). This compares to about 140
vascular plant species per 0.2 ha in temperate forests (Huston, 1993).

Soil samples were taken from the upper 8-10 cm of the umbric A horizon after
removing the overlaying organic root and litter layer, and placed immediately on wet ice.
After 48 h samples were shipped on dry ice to our laboratory in Michigan and stored at
-20°C. The soil moisture content was determined by drying soil overnight at 100°C.
Soil mechanical and chemical analyses were done by the Soil Analysis Laboratory,
Michigan State University, using the methods described in Peck et al. (1988).

DNA extraction and puriﬁcation. DNA was extracted from 10 g of soil using
the direct lysis method of Holben (1994) with the modiﬁcations previously described
for the Hawaiian soils (Chapter 2). Extraction efﬁciency was determined by comparing
the amount of DNA extracted with the amount of DNA expected calculated from the
difference in direct microscopic counts of bacterial cells before and after lysis. DNA
was quantiﬁed by ﬂuorometry (Chapter 2; Glasel, 1995). The amount of extracellular
DNA was determined following extraction with sodium phosphate before lysis (Chapter
4).

G+C fractionation technique. DNA fragments were separated according
to G+C content following the procedure of Holben and Harris (1995). Brieﬂy, DNA
was mixed with the A-T speciﬁc dye bis-benzimidazole, and the bouyant density of the
resulting DNA-bisbenzimide complex is decreased in proportion to the amount of dye
bound. A gradient of %G+C concentration was then established by equilibrium density

gradient ultracentrifugation. The amount of DNA in each fraction was determined by

39

absorption spectroscopy. DNA fractions were puriﬁed from the CsCl and the dye as
previously described (Chapter 2) to allow PCR ampliﬁcation. To verify the reliability
of the G+C ﬁ'actionation curves, repetitive analysis was done with two samples from
the same soil sampling bag.

rRNA analysis. Small subunit rRNA (SSU rRNA) genes were ampliﬁed
by PCR from puriﬁed soil DNA using the eubacteria speciﬁc primers fDl and rP2 of
Weisburg et al. (1991). PCR reactions were performed with Taq DNA polymerase
(Boehringer Mannheim, Indianapolis, Ind.) according to the manufacturer's protocol and
added quality control steps to ensure only soil bacterial rRNA genes were ampliﬁed
(Chapter 2). Ampliﬁed products were separated in agarose gels, and the bands
visualized by UV excitation after ethidium bromide staining. Prior to cloning, the
ampliﬁed SSU rDNA fragments were puriﬁed by spin column chromatography
(WizardTM PCR Miniprep; Promega, Madison, Wis), and an equimolar amount of
ampliﬁed PCR products was ligated to the vector pCRTM II (Invitrogen Corp., Carlsbad,
Cal.) and transformed into Escherichia coli Top-10F ’ competent cells. A primer pair
speciﬁcally designed to complement the polylinker of the vector pCRTM II was used to
amplify plasmid inserts directly from the transformant cells for SSU rDNA gene
screening (Zhou etal., 1997). To screen for SSU rDNA diversity, the ampliﬁed inserts
were digested with two sets of tetrameric restriction endonucleases, HaeIII and Hhal,
MspI and RsaI (Chapter 2). The resulting fragments were electrophoretically resolved
and the similarities between the electrophoretic patterns were analyzed using

GelComparTM (Applied Mathematics, Belgium)- cluster analyses by comparative

40

numerical analysis using UPGMA (Unweighted pair-group method using arithmetic
averages). Individual clones were grouped using a cut off of 97% similarity and a 5%
error rate for the band position. Each different pattern is termed a phylotype.

Determination of nucleotide sequences and phylogenetic analysis.
Puriﬁed SSU rDNA clone inserts were sequenced in both directions with the ﬂuorescent
DiDeoxyTM termination method using automated ﬂuorescent Taq cycle sequencing on
the ABI Catalyst 800 and ABI 373A sequencing system (Applied Biosystems, Forester
City, Calif). The forward primer was LFPl (Weisburg et al., 1991), the reverse primer
1529R (Weisburg et al., 1991) which targets a conserved region (positions 515-529 in E.
coli numbering). All sequences were aligned, and phylogenetic relationships were
inferred (Chapter 2). Only unambiguously aligned nucleotide positions were used for
the sequence analysis. Potential chimeric artifacts were evaluated by
CHECK_CHIMERA (Maidak et al., 1996), and mglobalCHI (Komatsoulis and
Waterman, 1997). Furthermore, any ARDRA restriction pattern seen more than once is

strong evidence against that clone pattern being from a chimera.

41

RESULTS

Changes in soil characteristics. The shift in vegetative cover to pasture
caused some changes in soil properties including a decrease in soil acidity and organic
carbon, and an increase in bulk density (Tablel). The higher cation exchange capacity in
the forest soil is likely due to its higher soil organic matter content.

G+C profiles. The G+C proﬁle of soil DNA was shifted to signiﬁcantly higher
G+C DNA with the change from forest to pasture (Fig. 1). The majority of soil DNA
was in the 55% to 70% G+C range for both soils but the major peak was 61% G+C for
the foreSt soil DNA and 67% G+C for the pasture soil. The pasture soil community
exhibited an additional peak in the 42 to 45% G+C range, and displayed a small peak of
71% G+C. DNA extracted from replicate soil samples showed repeated proﬁles such
that within site variation was much less than differences between the two sites. The
standard deviations of the mean curve differences for the forest and pasture soil
communities were small, 5 = 0.14 and s = 0.27, respectively (Fig. 1).

Phylotype replication within and between sites. The same phylotypes were
observed from replicate soil samples for the three dominant phylotypes in seven of the
nine cases for the forest and the pasture soils (Fig.2, dots). Examples of ARDRA data
documenting the same phylotypes in replicates of the pasture soil as well as different
phylotyes are shown in Figure 3. There was no overlap in dominant phylotypes
between forest and pasture samples, suggesting that the populations were different

between the two sites. Non dominant phylotypes were very different even between

42

Table 1. Differences in soil properties between the Kohala forest and pasture soils

 

 

Soil parameter Forest soil Pasture soil
Soil classiﬁcation Typic Placandept Typic Placandept
pH 4.2 5.2
Total carbon (%) 13.5 8.1
Total nitrogen (%) 1.5 ND
NO3'-N (pg/g) 2.2 5.6
NHf-N (pg/g) 83.3 22.4
CEC (meq/ 100 g) 89.3 51.9
Soil bulk density (g/cm3) 0.498“ 0.587
Dominating plant Metrosideros Pennisetum

polymorpha" clandestinum‘

 

’3 Data from Crews et al., 1995
b the overstory is dominated by Metrosideros polymorpha, a native C-3 tree
c the plant Pennisetum clandestinum is an African C-4 grass

42

43

Esosm 03 03m 088 05 “a BEES 88:3.— an 868:8
<23 mo 858m .owm mac» cm 23me 9 3:968 we? :8 Engage 05 v5 «88?? a $538 :8
25 .33 223— 05 5 28m “snowmen 95 Set 38838 <29 mo 8an 0+0 mo cemtaaﬁoo A 0.5!..—

<ZQ n6 0+0 es

 

 

 

on so on
. new." we ‘4.
x 5.. - . ..
. . %
o
I...
O m mob m.
< no: :8 580m s m
N
V
m not - e
. < not :8 83mg

 

 

 

44

 

 

 

 

A 401 63% G+C fraction
301

“a" 40 3 E

5. 1

o 20 ~

'3", 3o . '

.1: 1.

th 10 ,

3 20 4 Pasture

a.

’3; ' 0 a ' ' ' ‘ .

30/10- 1 10 20 3o 40 50 60

t:

3 Forest

U 0 ﬁr I H

 

 

 

10 20 30 40 50 60

H
1

 

 

 

 

 

 

   
 

Phylotype
B 25 “r °
, 35%G+C fraction

20 -
(D
a. - 1 -
>5 25 5
*5 n
”>1 20 « 10 ‘
é
H 15 - 5 ‘ Pasture
.. 11
Q" 10.1 0 r 1 r 1 T 1
7; L 1 10 20 3o 40 50 60
To“ 5 -
g Forest
a O _ I””IIIIIIIIIIIIIIIIIIlllllllllllllllllllIlllllllllllllll

 

    

1 10 20 30 40 50 60

Phylotype
Figure 2. Frequency distribution of ssu rdna gene phylotypes from the 63% G+C
fractions (A) and the 35% G+C fractions (B) ﬁ'om the rainforest and pasture
soils. Each graph shows the median and range (bars) of data of two replicates.
Solid dots (0) are placed above the phylotype which were identical in pattern
and rank for the two replicates. A replicate of the 35% G+C pasture sample
was not available.

45

replicates of the same vegetation type which lends support to the assumption that the
establishment of a clone library is a random rather than a biased process. Patterns of
rare clones (n=l) within the replicates of the 63% G+C pasture community overlapped
by 22% while the rare clones of the corresponding forest community replicates
overlapped by 12%. Only seven of the rare phylotypes overlapped between the two
vegetation types. The two most dominant patterns of the 35% G+C fraction of the
forest had very low frequencies of only 6.5(i2.5)% and 4.5(i0.5)%, but were found in
both replicates. There was only 4% overlap for the remaining patterns within the forest
replicates.

The phylotype abundance curves (Fig.2) of both the forest and pasture soil
replicates could be repeated successfully with very low standard deviation. The two
proﬁles of the 63% fraction of the forest soil were similar, i. e. 94% of all data fell within
a single standard deviation range of i1.17 of the mean curve difference. The
corresponding numbers for the pasture proﬁles at 63% were 93% at $0.77, and for the
35% G+C forest samples 98% at i061.

Changes in the ARDRA community profile. Microbial rDNA diversity in
the 63% G+C fraction differed between the pasture and the forest soil. The phylotype
and nucleotide sequence analysis showed that the dominant strains had shifted to
different, unrelated taxa. The rank abundance proﬁles of the 63% G+C ﬁactions are
separated by the stronger trend for dominance in the forest soil, but the much greater
phylotype richness in the pasture soil (Fig. 2). The four most dominant members

contribute 77%(i6) of the forest soil community while they represent about half

985 E:28o§e&V.E:3¢nE- ”commngsm mam ..

qaouwnnm otmﬁeﬁazm ”comﬂzwnsm and o

9.90 uueeES‘Qohzam. ”commngsm ADM e

aoﬁngsm Eatmteaehmow ”commE—Ezm mad 0

29:98 05 c8 won: 203 $8me 3:me bmsoswﬁﬁncs 3:0 9
Hana—$8 mono—o Mo 89:3: :33 05 .3 Rube—Ea 05 9

wﬁwao—op 8:20 .«o Spas: 05 wEEEV .3 35:28 6903.23 a 3 maﬁa—on mono—o 05 mo nonsense 0333M s

 

 

36 embeeewmess assessesz 2&5 2d? a m 3-2:
NS eeszeseessa eatseaeﬁsm 2&5 35 2 m 3.2:
:5 essaeEeessa geeseﬁsm 2&3 son a m 3.3:
n: esseseemsxsesesieem 295 932 w < aims:
one esszesesssﬁ amasseﬁsm 2&3 sum _ _ < Biz
2: esezesxeessa ateseﬁsm 2&3 sum a < 49mm:
=cm ohﬂumah
-1er ............. .1 mm wtzqewﬂme ........ am ammttsmsmmsm ........ e1 {ailimllldmwww .....
oﬁo e E53338 EzthucoEoV 823m $83633 M: m emtmmm
a? .802 a... see 525 assesses: we m $.95
0. Se anmztmﬁemm Ea~zoeﬁo~o§§mm~ own—panama agoEo&Qoh=Aw o < O©1mmm
v.3 eEESESeo Eztﬁogsﬁov. 823m $8393: 2 < mmﬁMI
NE a 02 .5 see 8255 assesses a. < £195
:3 3.0.8..—
emssne Ex 5 assuage g .eeeeeeene
Eta—Emﬁe ~53me @328 3330 $02 awesome—Em 9530M Baum—mom 0930793

 

 

.Ammzv :8 838m 05 was GE: :8 828:me 05 mo 8289a 0+9 $3 2: 80¢ $8232 95

2: mo mambo—Ea 28588 “88 8:: 05 .«o macaw <29 3mm no woman mecca—Ea Eczema—bi .N 935—.

47

(47%[i3]) of the pasture soil community. With 62(i4) phylotypes the pasture had a
phylotype richness which was almost 2.2(i0.4) times higher than for the forest soil.
The difference in abundance within the four most dominant phylotypes is more
prominent in the forest soil community. Clones of the ﬁrst phylotype are almost three
times as abundant as the next group while the respective difference in the pasture clones
is under 60% (Fig. 2A).

The pasture soil displayed a prominent dominance structure of the ﬁrst three
phylotypes from the 35% G+C fraction (Fig. 2B), while both replicates of the forest
soil show no signiﬁcant dominance pattern with no phylotype frequency over 10%.
The three most dominant members represent 47% of the pasture community, and only
about one third (15%[2t3]) of the forest soil community. The phylotype richness of the
35 % G+C fraction was about 43%(i4) higher than for the forest soil.

rDNA sequence analysis. One clone of each of the three dominant phylotypes
was sequenced. Most clones sequenced were only distantly related to known sequences
(Table 2). The dominant clones from the two vegetation types were very different
representing two very different phyla, F ibrobacter and Proteobacter. Many of the
phyla within the Fibrobacter group were dissimilar from each other (Table 3) while
those within the Proteobacteria were afﬁliated with Burkholderia and more similar to

each other (Tables 2 and 3).

 

bp
587
434
213
184

124

89
80

 

Figure 3. An ARDRA proﬁle following double digestion with HaeIII and HhaI of
clones form the 63% fraction of the pasture soil. Similar patterns can be seen
for three groups of clones, each found in both replicates: lanes 2,5,15;
6,7,8,10,l4,l7, and 9,11,12. Lanes 1 to 8 are from replicate A and lanes 9 to
18 from replicate B. Plasmid pBR322 digested with HaeIII was used as a size
marker (M).

49

Table 3. Evolutionary distances among SSU rDNA clones from dominant members of
the 63%G+C soil bacterial community from the rainforest and pasture soils.
The numbers represent percent SSU rDNA nucleotide sequence similarity.

A. F ibrobacter /Acidobacterium group

 

Clonea Evolutionary distance
for clone number”:

 

l 2 3 4 5

 

1.HRS-46 -

2. HRS-47 89.6 -

3. HRS-50 90.0 82.3 -

4. HRS-55 75.7 70.0 78.5 -

5. HRS-56 77.6 72.2 79.2 98.7 -

 

B. Burkholderia group

 

Clonea Evolutionary distance
for clone nmnber“:

 

 

1. HPS-45 -

2. BPS-52 99.6 -

3. HPS-54 98.0 99.0 -

4. HPS-6l 97.0 97.0 97.2 -

 

a Organisms are represented by respective

numbers in header.

 

detect
med

C Olllll‘.

heme:
analyz
Shows.
sites. ]
soil m

COR‘mL

inCTCESc
bacteria;
I930: (

lhe rel-”lat

50

DISCUSSION

The high microbial diversity even in this young, isolated Hawaiian soil makes the
detection and analysis of community components challenging. The molecular methods
used here, however, were able to distinguish clear differences in the microbial
community between the forest and pasture soil.

Because of the high microbial diversity in soil, distinguishing statistically
between inter versus intra sites is often very difﬁcult. In this study replicate samples
analyzed by G+C proﬁle, ARDRA abundance patterns and rDNA sequence analysis all
showed greater reproducible intra site results and signiﬁcantly different results between
sites. Hence, these methods may generally be useful to distinguish differences between
soil microbial communities. It is also important that these methods analyze the
community at different levels of resolution and different levels of comprehensiveness.
The G+C method provides an analysis at a coarser level of taxon distinction but is
comprehensive for all DNA, while the rDNA analysis is effective to nearly species level
but can only be done for ampliﬁable and selected dominant members. Hence, together
they form a complementary suite to better analyze community differences.

The change in vegetation type altered bacterial community composition and
increased diversity at least as revealed by the methods used here. The shift in the soil
bacterial community could be a result of either the disturbance of the clear cutting in the
1920’s, or the switch from forest to pasture, or both. A radical change in vegetation like

the replacement of a rainforest with a pasture certainly constitutes a major shift of the

 

soil en

ivy-1.3
. . »
Aldo:

r)

liable
em]. ‘
forests
regent
comm
Hawaii
4). lb
butnot
in the 6
dﬁlelop
thechai
not dis
gill-QUE]
ham-0n.
Commu.
.‘iOUng l
dEKeIOp
UllUSUal
COmPEIii

ofph} lo

5]

soil environment, especially in carbon sources for the microbes, and brings along a
change in soil properties like a decrease in soil acidity and an increase in bulk density
(Table 1) similar to those found in other ecological studies (Reiners et al., 1994; Spaans
et al., 1989). Furthermore, cattle graze in the pasture site while wild pigs disturb the
forest soil site. Hence there are both primary and secondary factors associated with the
vegetation shift that could be important determinants of the current microbial
communities. For my study of the chronosequence of similar soils on the big island of
Hawaii, I found that strong rDNA dominance patterns developed with soil age (Chapter
4). This dominance pattern was found in the 150,000 year old forest soil investigated,
but not in the 20,000 year or younger forest soils. However, the pattern of dominance
in the 63% G+C fraction of the pasture, which was established only eighty years ago,
developed a shape similar to the forest community (Fig. 2A). I therefore conclude that
the change in vegetation is most likely the factor responsible for the community to shift,
not disturbance. The dominance pattern in these fractions retain a similar shape,
although the decline in the ﬁrst seven phylotypes of the pasture community is more
harmonic than the respective group from the forest proﬁle. Perhaps the older forest
community has a more established system of specialists with deﬁned niches, while the
young pasture soil has a successional competition proﬁle which is typical for lesser
developed systems. However, the number of rare members in the pasture community is
unusually high compared to macroecological systems, which indicates no or limited
competition among soil bacteria (Zhou et al. , 1997). The signiﬁcantly increased number

of phylotypes in the 63% G+C fraction of the pasture soil suggests that the change in

52

vegetation type limited competition between the community members by opening more
niches to an already highly diverse community (Fig. 2A). Hence, I reject the assumption
of successional competition.

This proposed change in niches conceivably caused the observed change of
dominant phyla as inferred from rDNA sequences in the 63% G+C fraction. The
dominant taxa were from the F ibrobacter phylum in the forest soil and from the
Proteobacteria in the pasture soil. A statement about the change in the phylotypes
amongst the rare members cannot be made due to the extremely high bacterial diversity
found in soils. I assume a random selection as the underlying principle of our clone
libraries, hence any mixture of soil bacterial SSU rRNA clones can make up this
selection. A supporting argument to the niche addition theory is that community
diversity increases even though the relative amount of contributing community DNA is
lowered from the forest to the pasture soil (Fig. 1).

The application of molecular tools to analyze the soil bacterial community
described here enabled us not only to look at bacterial community proﬁles at several
levels of resolution, but also to investigate the dominant as well as some rare members
within these communities. The results support the common notion that terrestrial
environments exhibit enormous diversity, and contain novel, uncultivated organisms. A
more reﬁned picture of the phylotype abundance distribution within this community
could be gained by complete sequence analysis of all SSU rDNA inserts in the clone
libraries, but this is too costly. Furthermore, potential biases would always limit the
interpretation. To minimize bias, I have used the best available methods at present as

critically as possible.

CI-IAP'I'ER4

SOIL MICROBIAL COMMUNITY DEVELOPMENT
AS INFLUENCED BY SOIL AGE

ABSTRACT

Hawaiian soil sites were studied since these soils and ecosystems were reset by
volcanic eruptions to early stages of development, and because these soils of ages 200 to
150,000 years are underlaid by similar substrata, and inﬂuenced by similar climates and
biotic communities. Total microbial community DNA was extracted and fractionated
based on its G+C content. G+C proﬁles were similar for soils of different age except for
a slight decrease in low G+C content DNA in the oldest two soils. SSU rDNA gene
libraries were prepared from the 63% G+C fraction, which represents a dominant
biomass fraction, and from a 35% G+C fraction, which represents a minor biomass
fraction. The two major changes in community structure along the chronosequence are
the development of a dominance structure, and the decrease in phylotype diversity from
the youngest soils (200, 2,100, 20,000 yrs) to the oldest site (150,000 yrs). Maximum
phylotype richness within the dominant biomass fraction occurs at the mid-aged

Laupahoehoe site (20,000 yrs), which was conﬁrmed by both the maximum contribution

53

54

of high biomass DNA in the total DNA community proﬁle at this site and the maximum
in phylotype richness in the 63% G+C rank-abundance pattern which was even more
evident by their rarefaction curves. This maximal microbial diversity at intermediate soil
age is correlated with extreme values in several soil physical and chemical parameters,
especially plant available nitrogen, and soil bulk density. The ﬁndings are consistent
with the macroecological principle of diversity responses to nutrient availability. All
major changes with age described occurred in the fraction of high biomass (63% G+C),
whereas no statistically signiﬁcant changes could be found in the 35% G+C fraction.
Dominant phylotypes were found again in repeated experiments of the same soil
sample. However, dominant phylotypes were generally different between each site
along the chronosequence. These results indicate that variability within a soil site of a
particular age is not greater than between different soils. All of the soil clones sequenced
afﬁliate with the Eubacteria as was expected due to the speciﬁcity of the primer pair
used. None of the nucleotide sequences analyzed in this study matched known
sequences from any cultured species, or were identical to sequences in the databases.
Several of the most dominant clones belong to the clone cluster Vl (F ibrobacter phylum)
deﬁned by Stackebrandt et al. (1993), which forms an individual line of descent and is
now part of a new phylogenetic division. This novel group was more frequent at sites

older than 2,100 yrs.

55

INTRODUCTION

Soil microbial diversity is thought to be extremely high even at a small scale
(Torsvik et al., 1990). However, the magnitude of this diversity remains unknown
(Liesack and Stackebrandt, 1992; Bomeman et al. 1997). In addition, very little is
known about the environmental parameters that link ecological processes to that
diversity, and could inﬂuence the composition of the microbial community. Although
important microbial processes might occur at the small scale of microbial communities,
the great abundance of these communities in the soil extends their importance to the
larger ecosystem scale. Conversely, changes in parameters important at the ecosystem
level (e. g. , soil age, temperature, moisture, vegetation) may have signiﬁcant inﬂuences on
microbial communities, in particular on community structure and composition. These
microbial community changes may in turn lead to changes in diversity and activity of
larger organisms. Without insight into soil biodiversity, models of ecosystem
functioning and prediction of human impacts on changing ecosystem and global patterns
may not be understood.

Recent studies of soil bacterial diversity not only conﬁrmed the enormous
variety of microbes in soil but also corroborated the limitations of common diversity
studies. Use of the entire soil bacterial DNA as the target of study will only reveal the
most well represented species (Liesack et al., 1992; Bomeman et al., 1996).
Macroecologocial studies have long focused on certain groups of organisms, and not the

entire macroecological community. Consistent with this practice, I limited the target

56

group in this microeoological project to only parts of the soil bacterial community,
namely a fraction of soil extracted DNA with high biomass, and a fraction of low
biomass. Since cultivating methods only recover about 1% of all soil microorganisms,
molecular approaches were employed because they more comprehensively sample the
soil community and identify its most dominant members.

The DNA was separated according to its guanine and cytosine content (%G+C)
which provides a coarse level of separation of prokaryotic taxa (Holben and Harris,
1995). The fraction of high bacterial biomass selected had 63% G+C, and the fraction of
low biomass had 35% G+C. These DNA fractions were used as a template to generate
gene libraries of small subunit ribosomal DNA (SSU rDNA) by PCR ampliﬁcation.
Community structure was established by determining the diversity and abundance of
SSU rDNA clone types which were distinguished by digestion with two pairs of
tetrameric restriction endonucleases. Because the SSU rRNA gene contains diagnostic
variable regions that are unique to particular organisms, the digestion of the ampliﬁed
SSU rDNA gene with a pair of tetrameric restriction endonucleases results in mostly
unique restriction length ﬁagment polymorphism (RFLP) patterns. Since these RFLP
patterns are indicative of the phylogeny of the SSU rDNA gene, they can be used as
operational taxonomic units. For this study we use the term phylotype, which is
deﬁned as the pattern resulting from RF LP analysis of the SSU rDNA.

The purpose of this research was to determine how patterns of soil bacterial
community structure and composition change with the age and development of soil. I

studied the soils on the Big Island of Hawaii since they are the youngest and most

57

geographically isolated soils and hence changes in microbial community development
might be more easily seen. The soil was built from periodic volcanic eruptions, each of
which resets soil and ecosystem development. I sampled the four youngest soils of a
recently described chronosequence across the Hawaiian Islands, 1'. e. soil ages ranging
from 200 years to 150,000 years (Crews et al., 1995). Most importantly, many other
environmental variables were almost constant for the sites sampled. The parent material
shows little or no chemical variation (Wright et al., 1987), the precipitation and the
temperature are nearly constant throughout the year, and the discontinuous climatic
effects of pleistocene glaciation are minimized (V itousek et al., 1997). All sites support
intact rain forest vegetation dominated by the same tree species. By choosing these
young and geographically isolated Islands as my study site, I attempted to reduce the
extent of microbial diversity to a more manageable level. Their -- compared to older
soils -- lesser developed ecosystems are reﬂected in a low level of macrospecies richness

(Crews et al., 1995).

MATERIALS & METHODS

Soil origin and soil sampling. The chronosequence consists of four soils
located on the Big Island of Hawaii, ranging in age of the volcanic parent material from
200 to 150,000 years (Table 1, Fig. 1). Since these soils are matched in elevation,

precipitation, slope position, and disturbance history they form a sequence that varies

58

mainly in one parameter, soil age (Crews et al., 1995). The parent material of all soils is
basaltic tephra, a volcanic ash deposit (Clague and Dalrymple, 1987). All of the sites

are on minimal slopes (<6%) near 1200 m elevation, with 16°C mean annual air

temperature with little seasonal variation (Department of Geography, University of
Hawaii, 1983). The sites currently receive approximately 2500 mm of precipitation
annually (Giambelluca et al., 1986), which is well distributed throughout the year due to
the relatively constant NE trade winds (Carlquist et al., 1980). Soil thickness remains
constant at about 1 m. Each site is located under undisturbed, intact montane rainforest
with an unusually low level of species richness. The overstory of all sites is dominated
by the most widespread tree on the Hawaiian islands (Kitayama et al., 1995),
Metrosideros polymorpha, the native tree which increases in height from the youngest
site to the mid-aged Laupahoehoe site, to decline by around 50% at the oldest site,
Kohala (Crews et al., 1995). The well developed understory vegetation is dominated by
native tree ferns, Cibotium spp. Other trees and shrubs that account for most of the
remaining cover are Cheirodendron, Ilex, Coprosoma and Vaccinium (Crews et al. 1995).
The exotic plant species Psidium cattleianum and Hedychium garnerianum were also
present in low abundance (Riley and Vitousek, 1995). The total number of vascular
plant species per 0.2 ha was lowest at the youngest site with 34 species, and increased
with soil age to 47, 48 and then 53 species, respectively (Crews et al., 1995). As is
common for these montane areas of high precipitation, no nitrogen-ﬁxing vascular plant-

microbe symbioses was found within any of the sites (Vitousek and Walker, 1989;

59

$3: .3 a» @520 .3 Sun :0 393 2682330 ._
.8238 mm 88.3% 2E.

503809.22 .88 8mm Ea Dom: 30383859“ .35 =m 28 05mm on. 203 cozﬁﬁﬁoi 3:55. :82 28 22822th 3:52. :32 .

 

 

one 28 :5 3. 2:. S S: 222.5an 2.5. 233:2..8 835 223.
two 2: go we :2. S E: 328522: 2.5 .2 83.3.2 25.8 8285.93
:5 3.. E .o 2% 3 mm 82 222.522: 2%... .m_.m2\.w~.2 83 2.20
$5 22 mod an _..m on 2: 222.2285 22;: 833.32 8m 8.22:.
A 523 22.2 22.2 :5
2 :23 0:8 3038 28:2: 23:8 .3: cosmoEmmso CEZV own :8
2:5 =8 28 z 2.

 

 

In 3530 83232 conga—m =om 28:80.2 28:52.. 0:» =om

35:82 @2953 m0 3:5qu :8 was 3229.283 25 .— «Bah.

6O

Crews et al. 1995). None of the sites is known to have been cleared by humans (Crews
et al., 1995).

The youngest site is located near Thurston Lava Tube on the east slope of
Kilauea Volcano within Volcano National Park (Fig. 1). The site is fenced to keep feral
pigs (Sus scrofa) out of the Park. This soil is a Hydric Dystrandept developed on
several tephra (volcanic ash) depositions ranging in age from 200 years at the top
mineral layer to 400 years, and totaling approximately 38 cm in depth (Soil Survey
Staff, 1973). The sample analyzed was from the 200 year old layer. The next older soil
is located in Ola’a Forest on the north slope of Kilauea volcano. This soil is a Typic
Hydrandept originating from tephra deposits about 2,100 yr old (Soil Survey Staff,
1973). The 20,000 yr old Laupahoehoe soil is also a Typic Hydrandept, and is located
on the north east slope of Mauna Kea (Soil Survey Staff, 1973). The 150,000 year old
soil is a Typic Placandept (Soil Survey Staff, 1973), and is from the east slope of
Kohala volcano.

I removed the litter and the ﬁrst centimeter of soil, and sampled the next 7.5 cm
of soil. The samples were collected in the perimeter of Crews et al. (1995) main study
site. Duplicate soil samples of 750 g each were packaged on site in sterile
polypropylene bags and immediately put on ice. The next day they were placed in dry
ice coolers and shipped by express mail to Michigan where they were stored at - 20°C.

The soil moisture content was determined by drying soil overnight at 100°C. Soil
mechanical and chemical analyses were done by the Soil Analysis Laboratory, Michigan

State University, using the methods described in Peck et a1. (1988). The two oldest

61

 

Kohala
150,000 yr

  
  
     
     
    
      
  

Laupahoehoe
20,000 yr

)
‘

Waimea.
\

— 20°

 
 

A
. auna Kea

   

Hilo
\ Ola ' a

/ 2,1 00 yr

. Mauna Loa
A
Kilauea

Hawaii ‘

Volcanoes

Natl. Park

Thurston
200 yr
— 19°

156° 155°

 

 

 

Figure 1. Map showing the location of the four soil sites of the chronosequence on the
Big Island of Hawaii, HI, USA.

62

soils, Kohala and Laupahoehoe were sampled in 1992 and in 1994. To verify low
variability of the G+C fractionation curves and of the clone library, repetitive analyses
were done with two samples of the same soil sampling bag for the 1992 Kohala samples,
and compared with the 1994 Kohala sample collected in the same area. To determine
spatial variation within a sampling site, three samples were collected at the Laupahoehoe
site that were 3 m apart from each other.

DNA extraction and purification. Soil microbial DNA was extracted from 10
g of soil using the direct lysis method of Holben (1994) with several modiﬁcations as
described in Chapter 2. Subsequent DNA puriﬁcation was modiﬁed to include agarose
gel puriﬁcation and a single microcolumn passage of the excised and melted gel piece.
Extraction efﬁciency was determined by comparing the amount of DNA extracted with
the amount of DNA expected calculated from the difference in direct microscopic counts
of bacterial cells before and after lysis. Bacterial cells were counted directly using
computer aided microscopic counting procedures (Zhou et al. 1996). For consistency,
all counts were obtained by a single investigator. Details of the methods in this
paragraph are described in the Material and Methods section of Chapter 2.

Extraction of extracellular DNA from soil. In order to determine the amount
of extracellular DNA, 5 g of fresh soil (wet weight) was sequentially extracted four times
with 200 ml of 0.12 M sodium phosphate buffer at pH 8.0. DNA was removed directly
from the resulting extract by ethanol precipitation, and subsequently puriﬁed by spin

columns with a pore size allowing a molecular weight cut off of over 100K (Schleicher

63

and Schuell). Ogram et al. (1987) showed that this extraction procedure removes 99.9%
of previously added, radio-labeled DNA.

G+C fractionation technique. DNA fragments were separated according to
G+C content following the procedure of Holben and Harris (1994). Brieﬂy, DNA was
mixed with the A-T speciﬁc binding dye bis-benzimidazole, and the bouyant density of
the resulting DNA-bisbenzimide complex decreases in proportion to the amount of dye
bound. A gradient of %G+C concentration was then established by equilibrium density
gradient ultracentrifugation, and fractionated. To make PCR ampliﬁcation possible
DNA fractions were puriﬁed as described in Chapter 2.

PCR amplification of small subunit rRNA genes from soil DNA. Small
subtmit rRNA (SSU rRNA) genes were ampliﬁed by PCR from puriﬁed soil DNA using
the eubacteria speciﬁc primers fDl and rP2 of Weisburg et al. (1991). Details are
described in the Material and Methods section of Chapter 2.

Analysis of SSU rDNA clone library, determination of nucleotide
sequences and phylogenetic analysis. The ampliﬁed SSU rDNA fragments were
puriﬁed, an equimolar amount was ligated to the vector pCRTM II (Invitrogen, San
Diego, Calif), and the vector was transformed into Escherichia coli Top-lOF’
competent cells (Invitrogen, San Diego, Calif). A primer pair speciﬁcally designed to
complement the polylinker of the vector pCRTM II (44) was used to amplify plasmid
inserts directly from the transformant cells for SSU rDNA gene screening. Clones with
identical restriction patterns were digested with two additional tetrameric endonuclease

restriction enzymes. Cluster analyses were performed by comparative numerical

64

analysis and rank-abundance distribution patterns were established. The ampliﬁed SSU
rDNA clone inserts of the three most dominant phylotypes in each clone library were
puriﬁed, and partially sequenced at their 5’ end using automated ﬂuorescent T aq cycle
sequencing. For the detection of potential chimeric artifacts partial sequences that were
around 400 bases long were submitted to the CHECK_CHIMERA program offered by
RDP (Maidak et al., 1996). All sequences were aligned, and phylogenetic relationships
were inferred. Details of these steps are described in the Material and Methods section

of Chapter 2.

RESULTS

DNA extraction. High efficiency in DNA extraction is crucial to the
representativeness of the underlying soil bacterial community. I calculated the percent
of cells lysed using DTAF [5-(4,6-dichlorotriazin—2-yl)amino ﬂuorescein)] stained cell
counts, since DTAF counts commonly yield lower bacterial counts than the
corresponding acridine orange count (Suzuki et al. , 1993). Acridine orange is more likely
to nonspeciﬁcally stain organic particles, and these tropical soils are rich in organic
matter (Table 1). After optimizing lysis techniques I achieved a mean DNA extraction
efficiency of 88(i16)% (Table 2) based on loss of direct microscopic counts. Free
(extracellular) DNA which could add bias to the depiction of the soil bacterial

community does not appear to inﬂuence the amount of total community DNA extracted

22. 22.5.28 2.. 282222 :8 .22
<75 w 2-2 x v on 9 8.22.9.3 2.228 (75 32332 32.285 22 83258 223.22: 22 832 22> <75 22898— a
9: x :29: 2802 .228 825 \ 22: 22.3 328 2829-: .3 203—830 .
83223 82 2322.0. 825 ”223302 2328a H 328 .325 2382: 3:232 .22 22 2232.26 2288222 ..
A2928 8 83.2 25 ”223302 23233 H 2228 2325
:8 .22 A33 .2323 be :32» 82 2822228 $32222 222.5 .829 <Nm E2622 258—822. 22 28232 22322220 .

 

65

 

3w 3.2 28 $22 .2 2 8.23.3 .22 232.: 2852 2202
SW 4.: 22 25K .2 a. 2.92.2 .22 82222 25.8 8282823
2 2: «.2 2.2% .2 a. 6.34.2 .22 2.21.23 o2.~ «.22
2. we .3 $23 .2 1222: .22 834.8 8m 8.22:
2.2 26 223 3.2 26 2

22» 828220 36 3?: 2.82825 0.322052 96 3 2.8V Anmﬂ EUV Cb
..> 22> <22 22> <22 22> <22 2.: :8 2.28322 .2582322 .2528 25

Ill iullll! ill]!!! 4|. ﬂ

 

220232 222222.532
22322222 .«o 26m 92m me 83:23 322325 .5.“ 2:2» <75 23 2022250 22: :8 .2228 88:2 .2228 235 .N 222,—.

66

by the direct lysis protocol (Table 3). Initial extraction of DNA prior to lysis in the
oldest and the youngest soil did not recover detectable extracellular DNA. An attempt
to amplify the extract by PCR yielded no product, even after puriﬁcation of the extract.
In agarose gels the fragment size composition of soil extracted DNA was predominantly
of high molecular weight (i. e. , located above the 23 Kb marker) with only a faint signal in
the <23 Kb size range (data not shown).

Total community DNA proﬁles. The relative abundance of DNA of different
base composition (% G+C) was used to produce proﬁles of the community
composition, as well as to fractionate the DNA. The % G+C community proﬁles for a
site are reproducible as judged by the proﬁles from three replicate samples from one site
(Fig. 2). The community as observed by the level of G+C resolution appeared to be
stable over two years since the 1992 and 1994 proﬁles were similar (the standard
deviations (s) of the mean curve were 5 = 0.14 for two replicates ﬁom the same soil
sampling bag collected in 1992, and 23 = 0.40 between the two 1992 samples and the
1994 sample). High-molecular weight DNA of a limited size range was observed by
agarose gel electrophoresis (data not shown) showing that the DNA separation in the
gradient was not subject to potential bias of the varying molecular weight of randomly
sheared DNA.

The G+C proﬁles of total soil DNA for all sites along the chronosequence is
depicted in Figure 3. The proﬁles were standardized against DNA mixtures of known
composition (Fig. 3, insert). Each proﬁle had a distinct maximum between 55% and

70% G+C with a maximum of 63% (i2) G+C. This structure is typical also for

67

  

:1. ,. - - Ammm: <-dom 392”. 3.32320

83— 239» N 02m :8 023m 05 3 8.22.23 33 U Baozaom .N23
.2 8.2223 033233 :8 023m 05 mo 32238.2 083 m 93 < 832.8% .339» 28.8 C 33 33:3—
2: 82¢ 83223 :8 82: 28 8an 0+0 22 2232—232 .N 9.2»:—

on U+00\o
--r-//I/L//-IF o0

illillllll,l-lr lllri ll: 2

$8: 0.82 220“. 3.28. I

VNCI lelol 10%

3mm: 2.32 290“. 3.28. 3 .2

 

 

Ill...» ll ll ll L .\
DI!"

68

Table 3. Comparison of extracellular and intracellular DNA extraction yields

Cell DNA“

       

Soil Free DNAa Total DNA”

 

(pg/g dW) (Hg/g dW) Gig/g dW)
Thurston ND 6.4 6.5
Ola’a 0.1 14.1 14.2
Laupahoehoe 0.1 1 l .4 l 1.3
Kohala ND 19.8 19.8

 

a Free DNA yield with extra cellular extraction protocol

b Total DNA yield with direct lysis protocol

° Cell DNA yield with direct lysis protocol after extra cellular
extraction protocol

d Not detectable at a detection limit of <0.05 11g nucleic acid/g dw

C011!

f0

lau

mm

mm]

COm‘

69

midwestem agricultural soils (Chapter 2: Fig. 1). The ratio of DNA with high G+C
content (55 to 70%) over DNA of low G+C content (30 to 55%) was approximately 2:1
for the two youngest soils, and increased to 4:1 and 3:1 for the two oldest sites,
Laupahoehoe and Kohala, respectively. Over the chronosequence the proﬁles displayed
increasingly smoother curves.

Community structure as revealed by analysis of phylotype abundance proﬁles
of eubacterial rDNA libraries did not reveal an underlying community structure for the
youngest or the oldest soil in the chronosequence (for youngest soil, see Chapter 2: Fig.
3A). Therefore, total soil DNA fractionation was used to select a simpler set of the
community for analysis of structure. To lower soil bacterial diversity to a more
manageable level I fractionated total community DNA by its G+C content (Fig. 3), and
then analyzed bacterial rDNA ampliﬁed from ﬁve combined fractions over the range 62
to 65% G+C. Again, few repeating phylotypes were seen, hence the diversity was too
high to see any structure (data not shown). Only when I studied the structure of single
fractions of the %G+C gradient, was a distinct rank-abundance distribution of
phylotypes seen within the clone library.

Phylotype abundance patterns. Two fractions of the %G+C based
community proﬁles were chosen for comparative community structure analysis along
the chronosequence. The 63% fraction was chosen because it was located within the
major peak of DNA abundance typically found in previous soils studies (Chapter 2: Fig.
l). The 35% fraction was chosen to represent a community within the low biomass

range of the soil extracted DNA (Fig. 3). After amplifying the SSU rDNA inserts

70

60508 $5 mo gages—=86 2: 5 com: 56
3:803 0365» .«o $56an 8505— oouﬁ mo 83x?
a mo sewage.“ “t8:— 6059680520 2: «a £8 So.“

05 Sec 36866 <75 mo 88:8 G+C mo sauna—Ema .m 953..—

      
  

 

VNCI [9101.30 °/o

 

 

 

 

 

 

 

a.» 89.8
02.92.!!!—

 

 

 

s1-
VNCI 1330.1.” %

 

 

 

 

 

 

 

 

 

 

 

VNCI [3101.30 %

0+6 o\o

VNG W101. 1° °/o

 

 

333 .8263
.aoo m
8&5;th .85

 

 

 

o v
VNCI l9101 10 %

NO

00

o

_

71

£30m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

21.23::

 

 

 

00285334

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cocoa—608520
05 5 26m 58 05 .5,“ 3an bis—:88 05 8 mow—HE 9.6.x. Engage mo 8:35.80 .9 95»...—

mh—O no???

 

0+0 #3 2 exam
05 $2 2 gamma
0+0 «OS 2 $2:

 

 

 

 

 

 

 

\

 

 

\

 

 

 

 

 

 

_______:____:_ ...... ___________:_:_

 

 

 

.Xb
o/o
0
ll,
m.
Xmm my?
C
N
a\om v
O
m.
9
+
0
$3. .0
J
0
“J
a.

$8 _

 

d.

d:

d<

72

directly from clones they were digested with restriction enzymes and sorted by cluster
analysis. Clones with similar restriction patterns were further differentiated when
digested with a second combination of two restriction enzymes. The phylotype proﬁle
was repeatable in replicate samples as shown for three samples of the 63% fraction for
the Laupahoehoe site (Fig. 5). Similar results were found with three samples from the
oldest site, Kohala (data not shown). These data suggest not only that the ﬁndings from
this method are repeatable since the two soil samples from the Kohala site show rather
identical phylotype proﬁles. They also indicate that soil microbial communities that
were sampled only three meters apart at the same site (Laupahoehoe) have very similar
dominance structures (Table 4).

Figures 6 and 7 show the phylotype proﬁles for the 63% and the 35% G+C
fractions of the chronosequence. Table 4 indicates the size of the clone libraries
examined, and the relative abundance of the three best represented clones as a measure of
dominance relative to the entire clone library. With increasing soil age the three most
dominant SSU rDNA patterns represent 32%, 33%, 24(i5)%, and 67(i4)% of the SSU
rDNA insert diversity within their respective 63% G+C fractions. The insert in Figure
6 illustrates the ratio between the three best represented clones and the remaining clones
within each clone library along the chronosequence. In the 35% fractions the three
dominant phylotypes constitute 20(i4.5)% of the total.

Rarefaction analysis. The quantitative effect of using a fraction of the
community DNA on the diversity of phylotypes was evaluated using rarefaction curves

(Simberloff, 1978; calculation see Appendix A). Rarefaction analysis is a technique for

73

we
as“; @858 as Snooﬁmmwmma
- o :ozaozqom .
035 com 3an 8.325% x5e a
29:8 :8
-28 05 mo m

 

mambo—ham
i 3 cm V a. an 3 a
- — - = _= _ ======== = = = . C
u .3 \§_ . g
m 90“ =-=========

 

.k%.\\\\\\\\\\\\\\\\\\\=.===.===

4 Q0“ =============

    

 

 

§§§ a
I .r
.\\\ -
\\

adKJOIKllJ
13d (8)30013

74

.8 Bomb wcmoaosdom 02820:: .3 vomeam mace—o 28:65 838:: poem—Hovcb .moxon
33% £33302 5 E385: Son .3 333?: 2a mambo—Ea un—mamm .oocoscomocoﬁo 2: 3 29:8
:8 some 8m mambo—Ea EmﬁEov H88 035 cam 85.5: 0:20 5 8902.93 was 8:20 .3 555:2 .9 05am.

tb».%.3_>.£a_ 32:. Z.J::—U-—C gun—9232 .9: nor-3..-

43% no .8 "was. “m .< .8 8:6 3.95.0. .
.bo>=oo%o._ .2983 «Exam 05 E mocha—Em Esp—EEO “woe 2008 use “m5 2:

9 Hos—BE 203 0388 8:85.923 2: E mambo—Em “Saw—Bu 80E 2:: «Eu vacuum 05. ._
.550 some Eoc 2.80:6 8o? bEsﬁﬁoo
some 5 mambo—Em “55826 52: BE“ 93 288 .65 2.5. £2509 9:3 05 5h 88 Ste
gem a was 35:88 $3 90 to So a maﬁa 395% 203 8:20 3:232: airman 5520

3 coax—£8 2o? 858w“; actor—amok ho mEouaq 0:223:83 05 50369 moE§_E_m 2:. .

 

75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v v m - o - n m 3% 3
4 v m - o - w a 3% EN
0 v a - : - : A Asa; .2
S on an oz S oz mm a. guess;
a m» S Q2 3 oz E am 38.
2.2.2.: 0+0 $3
a a m N m m o m 3% E
ﬂ ﬂ m a ._1_ 3% EN
xl W a 2 ﬂ :3 E
8 2 t 2 he mm a. 9. ages;
MK 8 an S on E E 2 Bob
.0 m < u m <
£83233“ £33233— mambo—En
320M 8:00.333 «.20 53.55. co 8:20
2.53:. n+0 $8

mambo—En was mono—o go 898:2 .v 03.5.

76

.oocoscomoaoio 05 mac—m 35: 9:20
:08 :25» 8:20 wag“: 2: was 2.0323
0+0 o\em© 2: mo 8:20 850852 “mop

   

 

035 of 5953 0:3 ESE .58 magnum—Qua 8 R 3933 a: m
025 we mama?“ 0.8 $8 “moo—o 95 05 8 R 8
mo 8an 05 .mEoEtoaxo Bwﬁm u x \\ .:::\....\..\,.».\.\._. x... ..
2a m=8 “mow—Bo» 25 05 mo :onHMo N x \
8an 2t. .oocoscumoaoﬁo 2: Fr
3. a
90 £8 Bow 05 mo 552$ 0+0 8
$8 05 How $an 855% . _ \\\\\\\\\.\.\v.\v.\v.\vv.\vv.\.\v.\. . .8. . . 9
x52 cube—Em .e 0.5»:— .~ a o o a N . \\x\.\\\\\\.\..\..\w..\v_..\__
a .m _ O a.
I I i I e a 8 8
\\\\\...-..:.-.....:.
\\\\\\\\\\\\\\\\\\.\.\v\..,v.\_\.\v.\.\v.\.\.\. . : .8 o— R
i 25.2 : \\ \\\\\\\__ o
oonoosmmamq \, R
e 8 WW. 9
Q.
8 .
8:8:
u h 0 o o . C m ﬂ o AN 223. énzax— «.20
a _ a a o v—

Q

R

8

9.

 

9dA101£qd 19d (s)9u013

2&th

 

.xé

e\emN

3%

{cm

$2:

seuolo 1ueugwop
39m; ;0 uonnqmuog

77

05 .3 38 So.“ 05 mo corona 0+6 {own 05 .335: H
Sm 8an 8523.? v1.8 cube—Em .h 95»;

ooaoswomocoﬁo 05 mac—a has:

      

\\\,\_\\ i

.» com

\\ .

2

. ...\\\\\\\\\\..... w. .\. .x...v...._ _ _ :
. «N :
.\

Ca

    

 

; 25.8
oonoonaasmq

,. .uw\.\\.\\\\ I;

      

.o.

2.30 some 555 8:20 wﬁqﬁﬁa 05 Ba 8 m 0930— .3 m
.8on 0+0 $3 06 .«o mono—o 353298 “mop cm 3 cm
085 05 5233 0:8 ”:85 6053389850 \.\.\.\.\.\.\.\..VV.\..x.v. . . . 5... . . .8 2

 

 

.\

  

\\\\.R
.-

    

 

c

  

O

I
O

 

.cN

 

   
 

   

 

ov cm

    

 

 

  

 

 

\\\\\\\\\.\.\.\.\.\.\. . . . ...\... . . . . . . : _
. \\\\\. \\\\\\\. .. ..... My. NNN: . . in... .. cm
i coo o2 .8.\\.....__._._E c
a a O
_ a M .2 .2
.a r...
’00 . OM
Ax.
0§ i.
O?

$5? $3
adﬁwlﬁqd 19d (s)au013

l>

Estimated Number of Phylotypes

Estimated Number of Phylotypes

Figure 8.

78

 

    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

70 ‘
- -A- - Kohala-Nonfract. Laupahoehoe (20,000 yr) , ' l
60 - -v- - Thurston-Nonfract. .... _ . .V
W -‘D- Laupahoehoe . - ‘ .V
50 J—O—Ola'a ,5' '1'
—0— Thurston _ f . r " °°°°°
+ Kohala ‘- ° ' v- '
40 r - ."v"' - ......
.A ' . Z'. 0
30 . , .vg. ' ....... ...
.1 . -
. j .V
20 , 5:... ..
10 . ............ ' ‘J ........... \
Kohala (150,000 yr)|
0 l 1 A 1 1 L
0 10 20 30 40 50 60 70
Number of Clones
60

 

 

+Kohaia Kohala (150.000\yiﬂ
50 A —D-Laupahoehoc -

-O—Thurston .x‘.
’O
. ’1'
.o" \

Ola'a (2,100 yr)

   
 
  

 

 

4
o T --o--0la'a

 

 

 

 

 

 

 

 

 

30 ' o
,0
I
20
10
0 1 1 1 l
0 10 20 30 40 50 60 70 80

Number of Clones

Rarefaction curves for the phylotypes found in the 63% G+C (A), and in

the 35% (B) G+C fractions along the chronosequence. The X-axis depicts
the ex cted number of phylotypes calculated from a random samp e of
indivi uals taken from the total po ulation of lI‘lalhylotypes. A comparison
with the rarefaction curve for p ylotypes om nonfractionated whole
community DNA is shown

79

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0
E 0.8 , £ .
0
z u
N
0- 0.6‘- g —*
‘ 0.84
(I) 0.61 0.75
d) 0.4"" "_ —‘ 073
c .
2 .
o
0.2-r- —— -—‘ *—* I *1
0.26
0.22
0.19
0.0 -
200 2,100 20,000 150,000

Soil Age

Figure 9. A simpliﬁed diversity index, deﬁned as the ratio of all phylotypes of
contributing clones in a particular clone library versus the total number of
clones in that clone library, is shown for the four soils. The value of each
sample is shown in each bar as well as the mean and the standard deviation

80

comparing the diversity of communities. Given the species-abundance distribution of a
soil SSU rDNA clone library, rarefaction provides estimates of the species richness of
sub-samples taken from it. When plotted against the subsample these estimates deﬁne
the rarefaction curve. In Figure 8A I compare the cumulative diversity of total soil
DNA and 63% G+C fractions thereof for the chronosequence. The high diversity
throughout all 35% G+C fractions is summarized in Figure SB. This analysis veriﬁed
that the unfractionated DNA contained too many phylotypes to reveal any community
structure but that fractionation of DNA by its G+C content did reduce the diversity to a
level at which structures of dominance could be detected. However, this analysis does
suggest that due to enormous diversity the phylotype sampling in these fractions is far
from complete.

Community structure. Soil community structure changes can be seen across
the chronosequence. The relative abundance of dominant community members in the
high %G+C range is signiﬁcantly expanded in the oldest soil (Fig. 6). The tendency to
higher dominance over time appeared simultaneously in two ways. While the number of
representative SSU rDNA clones per phylotype increased for the most dominant
organism, the total portion of phylotypes in each soil library decreased. The most
dominant clones appear in percent of total clones over time with 13% (200 years), 17%
(2,100 years), 20% (20,000 years), and 35% (150,000 years), while the number of
phylotypes for the respective soils decreased after an initial increase from 46
phylotypes (200 years) over 49 and 66 (i8, n=3) (20,000 years), to ﬁnally 17 (:2, n=3)

phylotypes for the oldest soil (Table 4). Microbial species richness is reduced with

8]
increasing soil age. While there is no statistically signiﬁcant change in microbial species
richness from the 200 year old Thurston to the 20,000 year old Laupahoehoe soil,
species richness fell by 67% for the oldest soil, the 150,000 year old Kohala site (Fig.
9).

Comparative statistical analysis of entire rank-abundance proﬁles revealed a
statistically signiﬁcant difference only between the oldest Kohala site and the three
younger sites (p<0.05) (Table 5; calculation see Appendix B: Table 3). Similar
statistical analysis showed no signiﬁcant difference between any of the four proﬁles of
the 35% G+C fractions.

The increase of dominant structures in the oldest soil, and the decrease of
phylotype richness are illustrated by a simpliﬁed diversity index, the ratio of
phylotypes over the contributing clones in Fig. 9. This simpliﬁed diversity index is
used to compare community dynamics graphically with other soil physical and chemical
parameters in ﬁgures below.

A detailed statistical analysis of the ARDRA derived rank-abundance patterns is
discussed in Appendix B. Test statistics for the Chi-Square and the Kruskal-Wallis
tests were calculated for all phylotypes of a clone library sorted into three ranking
groups (Appendix B: Fig.4). The ﬁrst group is represented by the dominant phylotype.
The second group consists of the second to ﬁfth best represented phylotypes, and the
third group covers the remaining phylotypes. The Chi-Square test indicated a
statistically signiﬁcant lack of homogeneity across the chronosequence for the three

ranking groups (x2=l 39.2; p<0.05). Similar calculations lead to a rejection of the null

82

Table 5. Statistical comparison of community patterns over time. Aﬁer the test results
for the Kruskal-Wallis statistic indicated nonhomogeneity between
communities of the chronosequence the Multiple Comparisons Procedure was
used to determine if the community structures are signiﬁcantly different from
each other (p<0.05). All clone libraries were grouped by ranks prior to the
test. Details see Appendix B.

 

 

 

Pairwise comparison ___Si_gniﬁcant difference
of soil sitesa 35% G+C 63% G+C

Thurston and Ola’a NO NO
Thurston and Laupahoehoe NO NO
Thurston and Kohala NO YES
Ola’a and Laupahoehoe NO NO
Ola’a and Kohala NO YES
Laupahoehoe and Kohala NO YES

 

a Sites are grouped by soil age starting with the youngest soil

83

hypothesis of homogeneity for the 35% G+C fraction, although a Chi-Square test result
of 19.4 for these fractions indicated much less difference between the four soil ages. A
graphical comparison of determined and expected values for this test is shown in
Appendix B: Figures 4 and 5 for the 63% G+C fractions of the chronosequence. The
high diversity contributed by rare members of each clone library (Appendix B: Fig. 4:
ranking group 6th-...) has a strong effect on the Chi-Square value when the dominance
pattern becomes more accentuated with increasing soil age (Appendix B: Fig. 5) and the
two oldest soils contribute the most with their extreme ratios of phylotype diversity
over clone number. In addition to the Chi-Square test I also calculated the Kruskal-
Wallis test for this chronosequence to determine which soil in the chronosequence has a
community structure different ﬁ'om the others. In the Multiple Comparisons
Procedure, a part of the Kruskal-Wallis test, I used all factorial combinations of the four
soil ages to determine if there is a statistically signiﬁcant difference between the
structures of any two of the soil bacterial communites (Table 5). The Multiple
Comparisons Procedure shows that the rank-abundance distribution of the Kohala clone
library at 63% G+C is signiﬁcantly different from the rank-abundance distributions of
all three younger soils (p<0.05). There is, however, no signiﬁcant difference amongst
the three younger soils. For the rank-abundance distribution of the 35% G+C fractions
the Multiple Comparisons Procedure shows no signiﬁcant difference amongst the four
soils of the chronosequence (p<0.05). However, at a coarser level of signiﬁcance of
p<0.2 the rank-abundance distribution of the Ola'a and the Laupahoehoe clone library

are signiﬁcantly different from the rank-abundance distributions of the oldest, the

84

Kohala soil. There is, however, no signiﬁcant difference amongst the three younger soils
and between the youngest and the oldest soil.

Phylotype overlap between sites. Most phylotypes are unique for each soil
age. When I studied three samples from the same soil in the chronosequence for their
overlapping restriction patterns I found that diversity within the same soil is much
smaller than between soils of the chronosequence. When focusing on the three most
dominant members of each soil an overlap of the same dominant restriction patterns
could only be found once in the 63% fractions, between the two oldest soils at
Laupahoehoe and Kohala (Table 4). However, the overlap of dominant restriction
patterns within repeated samples drawn ﬁ'om the same chronosequence soil showed for
Kohala soil samples that two out of three dominant members in the 63% fractions were
identical. A similar result was found for three replicate samples drawn from the 20,000
year old Laupahoehoe soil (Table 4). As expected, the overlap of rare restriction
patterns (1'. e. , those found only once within a clone library) was low, in the range of 5 to
12% per site. These results indicate that variability within a soil site of a particular age
is less than between different soils. Sampling distance had no effect on detecting
difference in dominance structures for an order of magnitude. Two of the three repeated
samples drawn from Kohala soil were taken from the same soil aggregate in one soil
sampling bag. The three samples taken for comparison from the 20,000 year old
Laupahoehoe soil were drawn at a distance of three meters apart from each other. This
result corroborates the aptness of the method for comparative studies of soil microbial

communities since differences between sites can be distinguished from those within

85

own—280%.... Qacﬁguegﬁzaw E23333 mam ._ own—nEommé 53.532 E03333. mam .

83333 §t2o303o< E33333 mam » coax—«SW mace—o .8 235:: :38 05 .3 cameo—Ea 35
563:55 3.928933%: ”=o_m_>_3=m ma 9 wEwco—on 8:03 we 23:5: 05 9.63% .3 nae—3.8
mannsm Bceﬁemamam E2£>€n=m mam . .omboanm a 9 wEwcoﬁn 8:20 05 ..o 8523a 2623— ..

map—mu §mk§99§§0kw<u§m£QNm~sx ”=O_m_>=v£=w mam a 10m umOuOmF—mam— Gammmgam 80¢ DEC—o Ensues—Katina .mm: a

%-

 

 

3. ..Exomztoﬁomm Eézouﬁooobzquigeo aunoﬁoxmoxsgm. o 3-3—: $33
0 a 2 .5 £55338 §t2§§§ 8%.: .2825 a 8.55 O 28:83:
8 .R u: .5 .20 earn $938.3 a. £105 «302

-lump..-1--..------u--u_d._w.z4.wimmo ..... .m aﬁﬁmﬁﬂﬁ----mullamwmwtmilrimmed-
S :2 a... assess 538.8833. 82%: 88325 M: 8-2: 5 28:86
me .8 o: .5 .25 as}: $8885 a. tame: $8M

-..num...Iuum§ww§m$mmﬁ§mw§u®w~m®mnmmommmmmeﬂemmmtmwwm Ilium nuunrwmwwﬂiiiﬁmmad-
we :2 a... sesame... 53.58383 82:: 8.8832 2 3mm: 2 28:86
E a u: a... .2... 52%: $8883 a. csmmm 223

- u . $11.3. mam §$smmau§wummmmmcu<ullamagmuwmﬂnouam : lum 1..-:meme uuuuuuuuuuuu
ﬂ .2 02 ...... is Esra 8.88.2 2. 8-2: a 88:86
a. .99. a .5 a. sagéé 828m 29.5 sea 8 max: 8.2883;

--IMel---n..----uy§ws§§ww§mw<lummmmamumaﬂmﬂe 11m -----qumwwlllll!
mm seawetb. meaosonamnm «Exam 293 «5:80 _ _ m _ .55
3 ....ﬁts agosoeasascﬁx «8.8m 295 at? m. N _ -2: 8.22?
$er 033 .959“

began oases. mom 2 2.22 586 Susan censomoaé 233 8523.. 2; =8

 

 

 

.=8 :08 no.“ 2283 0+0 $8 05 Set 08 muse—o =< EPA—«SN 05 8.2 tom: 89$ £8me 3:me EmsoswEEwS.
3:0 .86on .565— wmomBo 05 9 55:87. 05 mm :03 mm £88505 ﬂ cube—Ea some mo 02895? 2522 2:.
.65.: :8 $2858 camaamm we 85338520 a 88m mausoleum <29 Dmm .«o mecca—Em awesome—bi .e 925.

86

sites, at least in this young ecosystem.

Phylogenetic analyses. All of the soil clones sequenced afﬁliate with the
Eubacteria as expected due to the speciﬁcity of the primer pair used (Table 6). None of
the nucleotide sequences analyzed in this study matched known sequences from any
cultured species, or were identical to sequences in the databases. Sequence similarity of
partial SSU rDNA sequences to SSU rDNA sequences in the RDP database (Maidak et
al. 1996) varied from 41% to 98%.

I subdivided the chronosequence nucleotide sequences into two general
categories, those related to previously recognized bacterial groups (Olsen et al., 1994)
such as Proteobacteria, C lostridia, Bacillus, and Syntrophmonas, and those related to
groups of novel descent such as F ibrobacter. All of the sequences of the only 35%
G+C fraction analyzed (the youngest soil) were afﬁliated with the genus Clostridium
(Chapter 2: Table 1). Of the three most dominant clones from the 63% G+C fractions
of the chronosequence, only the youngest site did not have clones related to members of
the group F ibrobacter (Table 6). Most members of the Acidobacterium subdivision in
the F ibrobacter Phylum (Maidak et al., 1996) are known only from rDNA sequence
analyses. None of the nucleotide sequences analyzed in this study that belongs to the
F ibrobacter Phylum is closely related to (< 27%) any previously described cultured

species.

87

 

 

 

 

 

 

 

 

 

 

0.8 10
w A
g h ' a
* -l
2 __ NH4 8 n
o 0.6 c g
'- r:
\ m.— '—
w _ a) m
a) / 6 tit-g, 2 NE
D. N o m m x
a o" - / 2 E 0 g)
o 5 m
— d)
4 ,_ 3
E - / =3 e 3
CL 0 - 5 g
- Z
2% o 2 / 3 2 g
a)
CI _ I. I
III... Ilalllll1) I In? E
II
0 '—' . IvIv'] v u - -.','l ; o
100 .5. 1000 -: 100002
: 6 8
2 5
'- SOII Age %-

 

Figure 10. Plant available nitrogen and phosphorus determined in situ with anion-
exchange resin bags that were buried ﬁ'om 5 to 8 month (data from Crews et
al., 1995). The data are overlaid with the diversity index from Figure 9

 

88

DISCUSSION

Soil microorganisms play an integral role in the maintenance of soil fertility and
soil structure, and certain functions are performed exclusively by bacteria. Nevertheless,
many soil ecological studies tend to ignore that important third group of organisms with
a bias towards vegetative and animal life, or, at best, limit it to a black box of “microbial
activity”. Knowing and understanding soil microbial communities is essential for a
sustainable biosphere, from agricultural efﬁciency questions to biodegradation of
hazardous waste spills. However, achieving a comprehensive picture of a native
bacterial community represents a signiﬁcant task. The challenge becomes even greater
when one moves this investigation into one of the most complex environments, soil,
with its high species richness and evenness. Several recent studies have attempted to
describe a few of the most common members of a soil community. However, I believe it
is no longer enough to only inventory soil organisms, but we must also link that
diversity to ecological processes.

In macroecological systems the following relationship between productivity and
biodiversity is typical: In low-nutrient environments an increase in biodiversity seems
to enhance productivity (Odum, 1998). In high-nutrient environments an increase in
productivity increases dominance, and reduces diversity. This macroecological principle

seems to be consistent with the microbial community dynamics observed at this

89

3:258

 

 

 

 

 

 

     

 

0 0 0 0 0 O 0
7 6 5 4 3 2 cl
- _ u d _
s
e
w
T...
mm
a
M%
Cf
.mm
3
@b
Om
ou
o/N
E-
n n _ n n —
o o o o o o o
7 6 5 4 3 2 1

Eras. oEmQO o\o

2100 20000 150000

200

Soil Age (yr)

Figure 11. Comparison of soil organic matter content and phylotype

diversity for soils of the chronosequence.

 

    
  

  
  
 

 

 

 

2200 2
3 1.4 ~ 2
0
A ‘- In
q
3\: .5. 1-2 ” - 2100 5'
z m
g a FollarN ,_,.2
o a 1 ’ l “3: 1,
c2J g - 2000 4.3
g 0.8 " 1 2'
g a 0 6 Decomposition ....» NPP '-' '3’-
- ” .-
u. ,5 constant,» _ 1900 2
£- 0.4 ~ -"" ' %
v
V
0.2 V V YY'YUTI V l I YIIIVI T T YYYYWTI 1800

100 g 1000 . 10000 , 100000

‘3 3: g 2

5 ° 3 g

f f. x

o a.
Soul Age (log yr) 3

Figure 12. Exponential decomposition constants and foliar nitrogen concentration for
leaf litter of the dominant plant Metrosideros polymorpha. Decomposition
constants were determined over 2 years by in situ decomposition (data
from Crews et al., 1995). Foliar nitrogen concentration is calculated in %
dry mass (data ﬁ'om Vitousek et al., 1995). Net primary production is
shown (data from Herbert, 1995).

91

chronosequence. A relationship between soil productivity and microbial diversity can
be found when the three younger, less developed soils are compared with the oldest soil
at the Kohala site. In the younger soils, the nutrient supply has not yet reached its
maximum which is reﬂected by high phylotype diversity (Fig. 9). From the youngest
soil to the mid-aged Laupahoehoe site, plant available nitrogen and phosphorus are
increasing, with a nitrogen maximum at this age (Fig. 10). Between 2,100 years and
20,000 years of soil age the organic matter reaches a maximum (Fig. 11; Table 1). This
coincides with a slightly higher phylotype diversity for the dominant (63% G+C)
microbial fraction at the 20,000 year old Laupahoehoe site. The highest turnover rates
for NH4+ -, NO3' -, and N20 - nitrogen also occur at this site, which indicates maximal
microbial activity, and soil productivity (Fig. 10; Crews et al., 1995). After reaching
this high level of nutrient availability at the mid-age site the resulting high productivity
is followed by an increase in microbial dominance (63% fraction) at the expense of
phylotype diversity (Fig. 6) at the next older soil.

Specialization is offered as the most reasonable underlying mechanism for this
change in diversity at the oldest site. With time specialists for the decomposition of the
abundant accumulated organic material may take over the available niches in the soil.
This specialization would increase dominance at the expense of diversity, which is the
pattern at the Kohala site. This assumed specialization is reﬂected in the decomposition
dynamics along the chronosequence. Concentrations of elements in live tree leaves
generally reﬂect available nutrient pools in the soil, for both the dominant tree

Metrosideros polymorpha and for other tree and bush species across the chronosequence

92

(V itousek et al., 1995). While foliar nitrogen and phosphorus contents in Metrosideros
polymorpha leaves do not change, the leaf litter decomposition rate increases
signiﬁcantly from the midage to the oldest soil, and climaxes in a three-fold faster
decomposition rate compared to the youngest site (Fig. 12; Crews et al., 1995). Control
experiments showed that this increase in decomposition rate was site speciﬁc. Also,
after two years of decomposition almost four times as much leaf litter mass remained at
the mid-aged Laupahoehoe site relative to the oldest site, suggesting improved
decomposition ability of the microbial community at that site. This difference in
decomposition rate occurred despite the constancy of both precipitation and
temperature across the chronosequence. With time the soil microbial community might
develop and change to. a K-selection strategy at Kohala, where efﬁciently competing
decomposers operate at maximum decomposition rates.

This assumption is corroborated by the amount of nitrogen and phosphorus
remaining within partially decomposed leaf litter after a period of two years (Crews et
al., 1995). Almost four times as much nitrogen and phosphorus remained in the leaf
litter at the mid-aged Laupahoehoe site (34% of initial N and P) versus the oldest
Kohala site (9% of initial N and P).

Over the course of soil development along the chronosequence the amount of
plant available phosphorus is constantly increasing, while plant available nitrogen
peaked at the midage Laupahoehoe site, and then fell to less than 50% that value at the
oldest site (Fig. 10; Crews et al., 1995). The C/N ratio stayed almost the same at the

last two sites in the chronosequence suggesting that changes in plant available nutrients

93

or litter quality were not responsible for a change in microbial community structure, and
hence specialization within the community is hypothesized (Table 1).

The dominant soil minerals might also contribute to the high microbial activity at
the two oldest soils. Weathering processes are rapid both because of the wet tropical
climate and the reactive nature of the lava parent material. The rapidly weatherable
olivine, glass and plagioclase are completely consumed before 20,000 years (V itousek et
al., 1997). Subsequently formed noncrystalline minerals (allophane, ferrihydrite, and
imogolite) are characterized by large reactive surface areas that bind cations, phosphorus
and soil organic matter very effectively (Wada, 1989), which aids productivity through
accumulation of organic matter (Table 1; Fig. 11).

Specialization of the decomposing microbial population could have a mutualistic
effect on the vegetative cover. The rapid decomposition of leaf litter with relatively high
nutrient concentrations would constantly supply plants with high levels of available
nutrients (Fig. 12). This rapid regeneration of available nutrients provides a positive
feedback between the plants and the decomposing microbial community, in which high
levels of available nutrients are maintained (V itousek, 1982). Net primary production of
the forest increases steadily between 200 and 20,000 years, but the average rate of net
primary production exhibits a substantial increase after 20,000 years to peak at the
oldest site at a level 20% over that of the youngest site (Fig. 12) (Herbert, 1995;
Vitousek et al. , 1997), which is three times as high as found in midwestem forests.

Profiles of the fractionated community DNA. The G+C based fractionation

gradients represented with excellent repeatability community proﬁles characteristic of

94

the relative composition of the soil community. The similarity of the replicate gradient
proﬁles at Kohala (Fig. 2) demonstrated that these curves represent not a just
“snapshot” in time but are repeatable at the same soil site over time. Hence, differences
in proﬁles show the effects of major changes in the relative abundance of components of
the soil microbial community. A population change of at least 105 to 10‘5 cells is
required to cause a change in G+C peak height in a total community DNA proﬁle. The

majority (60% (i3)) of the extracted soil DNA corresponds to the 55% to 68% G+C

range which includes genera known to dominate soil bacterial communities, e. g.
Agrobacterium (55-61%G+C), Alcaligenes (56-63%G+C), Arthrobacter (63-69%G+C),
and Pseudomonas (58-66%G+C) (Alexander, 1977; Holben and Harris, 1995). Soil
DNA found in the 29% to 47 % G+C range corresponds to genera like Streptococcus
(35-4O %G+C), Clostridium (31-36 %G+C), and Bacillus (38-45 %G+C). Rarefaction
analysis indicated reduced diversity in the individual % G+C fractions compared to the
total DNA, especially for the 63% G+C series, but that rDNA diversity was far from
exhausted in all clone libraries [83(13) clones] of this ﬁaction (Fig. 8A).

Bacterial diversity in soil environments was too high to be analyzed
comparatively between different sites based on the direct sequencing approach. The
studies of soil community structure found to date tend to concentrate on the relatively
few and likely more abundant species, but do not recover the rich variety of organisms
that occur in smaller numbers but encompass a much greater range of species. These

microorganisms may be important in acting as a reservoir of different physiological

95

types, which may respond by population size increase following changes in the habitat.
This would also imply a degree of built in redundancy in the microbial community, and
that some organisms are remainders of past conditions. Knowledge of the composition
of this reservoir will allow us to distinguish between a complete community shift and a
change in only the dominant community members. Furthermore, evidence from 3H-
labeling experiments and G+C based ﬁactionation of soil community DNA indicates
that activity and growth are not conﬁned to a small ﬁ'action of the bacterial biomass but
are widespread among the genera represented in the proﬁle (Harris, 1994). Hence,
separation by G+C content uncovers new diversity and provides evidence that soil
rDNA diversity is much higher than can be detected by direct rDNA analysis of total
community DNA. This higher estimate of soil rDNA diversity is supportive of the high
bacterial diversity estimate reported by Torsvik et al. (1990) based on rates of soil DNA
reannealing.

DNA extraction. A signiﬁcant concern in establishing a community picture
from total community DNA extracted from an environment is whether cell lysis was
partial or biased. The pre- and post-extraction direct microscopic counts indicate
sufﬁcient lysis efﬁciency was achieved. The high diversity found in the phylogenetic
analysis of some of the rDNA clones together with ﬁnding a high number of sequences
for sporeformers conﬁrmed the reliability of my approach (Chapter 2: Fig.7). The
efﬁciency of total community DNA extraction is crucial for a comprehensive evaluation
of the soil microbial commrmity. Initial difficulties with DNA extraction (not reported)

might have been due to a parent-material effect. Weathered tephra, a Hawaiian volcanic

96

ash and the source of the parent material, is characterized by a high content of allophane
at soils older than 2,100 years. Allophanes are amorphic hydrated aluminum silicates
which are able to ﬁx dissolved phosphate at high rates (Nanzyo et al., 1993). For
maximum yields of soil DNA, I saturated active sites on allophane by increasing the
phosphate concentration in the extraction buffer one hundred times (to 100 mM
phosphate). Extracting DNA ﬁom Andosols is also made difﬁcult by their high content
of organic matter. The allophanic clays in these soils help to stabilize soil organic matter
such that soil organic matter contents two to three times higher that of nonallopharric
materials are formed (Paul and Clark, 1996). The abundance of humic substances in
soils, many of which coextract with nucleic acids in standard protocols, required post-
extractive puriﬁcation to allow for PCR ampliﬁcation of the extracted community DNA.

The potential for bias is also present once comprehensive lysis is achieved.
Biases can come from preferred ampliﬁcation in the PCR reaction, the formation of
chimeric SSU rDNA molecules, or the rrn copy number variation in different bacterial
species (Chapter 1: Background). These biases have been extensively reported or
studied elsewhere (Ward et a1. , 1992; Stackebrandt et al., 1993; More' et al., 1994; Ogram
et al., 1987; Tsai et al., 1991; Liesack et al., 1991; Suzuki et al., 1996; Robinson-Cox et
al., 1995; Reysenbach etal., 1992; Farrely et al., 1995; Kopczynski et al., 1994; Mead
et al., 1991; F inney, 1993). For a complete quantitative determination of a soil bacterial
community, many of the techniques are still in need of reﬁnement. I have used the at
present best available methods as critically as possible to minimize the introduction of

potential biases.

97

Phylogenetic analyses. In a recent comparative study of soil phylogenetic
surveys based on rDNA sequence studies Kuske et al. (1997) show considerable
commonality of sequences among the surveys that fall outside of the known bacterial
domains (Olsen et al., 1994). This remarkable abundance of members of this new
phylogenetically diverse group in clone libraries across soils from vastly different
geographic regions indicates that these organisms are dominant regardless of soil types,
geographic regions, and overlying vegetation (Kuske et al., 1997; Liesack et al., 1992;
Stackebrandt et al., 1993; Ueda et al., 1995; Bomeman et al., 1996; Bomeman et al.,
1997). My ﬁndings support this thesis since members of the dominant biomass (63%
G+C) of chronosequence soils analyzed also belong to the same novel branch (Table 6).
Interestingly, they are all related to the clone cluster VI deﬁned by Stackebrandt et al.
(1993), which forms an individual line of descent currently under analysis (Ludwig et al.,
1997; N. Pace, personal communication). Kuske et al. (1997) made the ﬁrst attempt to
compare SSU rDNA sequence data among various other soils including their own, and
discovered substantial similarities of phylogenetic types across different soils. Most of
those sequences fall into several diverse groups distinct from previously recognized
bacterial divisions. Of all nucleotide sequences analyzed in this study, only 40% are
related to previously recognized bacterial groups (Olsen et al., 1994). Of these, their
closest relatives are members of culturable genera found in many types of soils, such as
Proteobacteria, C lostridia, and Bacillus.

Comprehensive inventories? Figure 8 clearly demonstrated that

complete diversity has not been captured in any of the communities analyzed, since

98

none of the rarefaction curves level off with increasing clone numbers. Zhou et al. (ESA
abstract No. 226, 1997) attempted to analyze a continental soil clone library ﬂour 5 g pf
soil with over 600 clones, but found this was still not large enough to capture full
diversity, i.e. the respective rarefaction curve was still linear. The reason to avoid bigger
clone libraries is not only the prohibitive cost and labor, but also the ecological value of
such an attempt. No macroecological study attempts to gather all species of an
environment. Instead, studies are restricted to a certain group, usually a ﬁrnctional
group, which makes the analysis feasible. My attempt is analogous, namely to select
and evaluate a fraction of the enormous soil bacterial diversity. The ﬁeld needs to get
away from the notion that one can be comprehensive. It is simply not yet possible.

Limitations. The interpretation of these data is limited by the fact that there
are no comparable studies. Even though some of these data are based on triplicate
experiments, generalizations are difficult to make; for example, only two ﬁ‘actions of
total community DNA were investigated. All major changes described occurred in the
fraction of high biomass, whereas no signiﬁcant changes could be found in the G+C
fraction representing more minor members of the community. This does not exclude the
possibility that dominance within other high G+C fractions, e.g. 61% or 67% G+C
fraction, might have very different responses to soil age. Further investigations on this
soil series would be necessary to clarify this.

This chronosequence on the Island of Hawaii supplied a relatively simple
system of developing soils to study and evaluate microbial community structure and its

interaction with soil developmental processes over time. However, even though I know

99

that the Hawaiian islands are not recovering from the history of disturbances like the
Pleistocene glaciations in temperate regions, the history of disturbances in the sampled
area is not entirely known. Furthermore, I cannot claim long term stability or
equilibrium stability in these tr0pical soils, since there is a constant input of nutrients
and possibly microorganisms via the trade winds. Exogenous inputs of Asian dust
deposition (Jackson et al., 1971) might also be a constant source of nutrients and
nonindigenous microorganisms. The possible inﬂuence of these inputs may prevent
equilibrium in these communities. In addition, fungal activitiy might be the initial step
of foliar decomposition. Also, the encountered community transition rates cannot be
applied globally, since they probably occur more rapidly in Hawaii than in other soil
sites due to abundant rainfall, warm year-round temperatures, and easily weathered
primary minerals. Nevertheless, the understanding of community development gained
on this Hawaiian chronosequence is an important ﬁrst step to understanding
developmental changes of microbial communities, and might be broadly applicable to
continental or other island systems. The increase of microbial phylotype dominance
along the chronosequence, which was offset by a decrease in phylotype diversity after
the intermediate stage of soil development, rrright represent a typical development for

the major fraction of microbial biomass in other mature soils.

APPENDIX A

APPENDIX A

STATISTICAL COMPARISON OF COMMUNITIES
BY RAREFACTION CURVES

When analyzing community structure we need to know when we have sampled
enough organisms to sufﬁciently characterize the community. When applied to SSU
rDNA analyses of communities this equates to determining when a clone library is large
enough to characterize the sampled bacterial community, i. e. when diversity in the
‘number of species’ versus ‘number of individuals sampled‘ curve levels off (Gleason,
1922) (Fig. 1). With the rDNA method different taxa can be deﬁned as clones with
different restriction fragments derived form using several tetrameric endonucleases: these
different groups have been termed phylotypes. More speciﬁcally, such analyses
actually reﬂect the population distribution of ampliﬁable rDNA operons. While this
does not represent species in a formal sense, it is a measure that groups like individuals,

and hence “species” diversity measures can be used to characterize rrn populations.

100

101

>

 

 

3

D.

2:

O

3“ B

..G

0..

“-1

0

3'5

.0

E A

2

>
Number of Clones

Figure 1. Sampling plot of the dependence of diversity on the number of individuals.
Region A refers to the curve area of rapid accumulation of phylotypes associated with
increasing number of clones sampled. As the number of clones added becomes larger,
region B, the clone library has become large enough that new phylotypes are added at a
much slower rate (adapted after Sanders 1968).

The clone libraries collected ﬁom an environment are ﬁnite samples of a much
larger community. We can measure the phylotype abundance distribution within a clone
library, but of ecological interest is the diversity of the bacterial community as a whole.
To estimate the trend of species richness in a clone library with respect to the sampling
curve (Fig. 1), rarefaction analysis can be used. Furthermore, the rarefaction curves of
several communities plotted together render a graphical display of the differences in
diversity of those communities.

Rarefaction should be used only for clone libraries that have been obtained by using

standardized sampling and processing protocols. Rarefaction curves are not to be used

to extrapolate beyond the number of clones collected (Tipper, 1979).

102

Rarefaction analysis. Rarefaction analysis is a statistical technique for
comparing diversities of organismal communities. Given the species-abundance
distribution of a collection, 1'. e. a clone library, rarefaction confers estimates of the
species richness of sub—samples taken from it. When plotted against the subsample,
these estimates deﬁne the rarefaction curve, which is a deterministic transform of the
phylotype abundance distribution within the clone library. Rarefaction curves have
been calculated to compare phylotype diversity of total soil community DNA versus
fractions of this community DNA based on G+C content, and to compare these
fractions among different soils. Rarefaction allows for unequal sample size and
estimates the species richness (i.e. number of phylotypes), E(Sm), that a total
population of N individuals would have, if its size were randomly restricted (rareﬁed) to
a sample of m individuals (i.e. clones taken from a clone library) (Tipper, 1979).

Rarefaction is computed using the equation:

E<Sm>=éll-t”;”)/(le

where S is the number of species in the parent collection, N, is the number of individuals
of the i-th species, N is the number of individuals in the parent collection (i.e. the entire
clone library), and m is the number of individuals in the rareﬁed sample (m<N)
(Sirnberloff, 1978). Since this calculation has no probabilistic basis, Simberloff (1978)
created the software program SIM that calculates the variance of the expected species

richness E(Sm).

103

The SIM rarefaction program. The rarefaction program SIM is written in
FORTRAN-IV for a maximum of 550 operational taxonomic units. This program
computes the expected species richness (E(SN)), the variance Var(SN) of this estimate,
and the standard deviation (SD) of the estimate for rareﬁed, i.e. restricted, samples.
These values are needed to compare the phylotype richness of clone libraries of unequal
size. The data entry and output are shown in Figure 2; an example of rarefaction curves

calculated with the software SIM can be seen in Figure 3.

Entry form:

 

 

N NS NUM
89 67 15
CP. CPz CP3....CPns

18431111111111111
11111111111111]
111
llllllllllllllll
lll

nurm numz DUITI3... numNUM

1510152025303540455055606570

 

 

Output form:

104

 

N
89

NS
67

NUMBERS OF INDIVIDUALS IN THE DIFFERENT SPECIES

N

1

5
10
15
20
25
30
35
4O
45
50
55
6O
65
70

184

1
l
1
l

Had—n

E(S)
1.0014
4.6532
8.7892
12.7213
16.5618
20.3495
24.1009
27.8210
31.5156
35.1878
38.8413
42.4792
46.1044
49.7196
53.3271

311

SD.
.3963
.7028
1.1487

1.4472
1.6537
1.7871
1.9102
1.9898
2.0454
2.0714
2.0705
2.0391
1.9773
1.8793
1.7407

11111111111

VARIANCE
.1571
.4939

1.3195

2.0943
2.7348
3.1938
3.6490
3.9594
4.1836
4.2909
4.2869
4.1578
3.9097
3.5318
3.0300

 

 

Stop - Program terminated.

 

Figure 2. Example of data entry and output for the rarefaction program SIM using data
from a fraction at the youngest Hawaii site, Thurston.

The data entry (top part of Fig. 2) requires the the following raw data:
0 N, is the number of individual clones in the clone library (here: 89)
0 NS, is the number of phylotypes

NUM, is the number of subsamples, mm, of individual clones to be drawn from

the clone library

105

o numi‘ is a step in the array of subsamples in ascending order
0 CP, is the number of clones in each phylotype

0 NUMBERS OF INDIVIDUALS IN THE DIFFERENT SPECIES,

is the number of clones in each phylotype sorted by descending
abundance

70

 

m

 

—e—A: Total Soil DNA A
60 -

+C: 63% G+C Fraction
—l- -B: 35% G+C Fraction

 

 

 

 

 

 

0 r r r r r r r
0 10 20 30 40 50 60 70

 

Estimate of Phylotype Diversity, E(S )

Number of Clones, N

Figure 3. Rarefaction curves calculated by the rarefaction program SIM for comparison
of three related clone libraries. Non-fractionated DNA (A) is compared with fractions
of the same sample ﬁrst separated by its G+C content for their phylotype diversity.
Since rarefaction curves produce values of E(Sm) only for integer values of m, the
datapoints representing the rarefaction curve must be identiﬁed.

Rarefaction is a weak statistical technique, since identities of the species are not used.

Because the rarefaction curve is a mere transformation of the original phylotype-

106

Rarefaction is a weak statistical technique, since identities of the species are not used.
Because the rarefaction curve is a mere transformation of the original phylotype-
abundance distribution, I suggest here that comparison of rarefaction curves for other
than graphical reasons should more proﬁtably be replaced by direct comparison of those

distributions, using the Kruskal-Wallis test statistic (Appendix B).

ACKNOWLEDGMENTS
I thank J. Dunbar for allowing me to use his updated version of Sirnberloff‘s (1978)

FORTRAN program SIM.

REFERENCES

1. Gleason, H. A. 1922. On the relation between species and area. Ecology. 3: 1 58-162.

2. Hurlbert, 8.11. 1971. The non-concept of species diversity: a critique and
alternative parameters. Ecology 52:577-586.

3. Sanders, H. L. 1968. Marine bentlric diversity: a comparative study. Am. Nat.
102:243-282.

4. Simberloff, D. 1978. Use of rarefaction and related methods in ecology, p. 150-165.
In K. L. Dickson, J. Cairns, Jr., and R. J. Livingston (ed.), Biological Data in Water
Pollution Assessment: Quantitative and Statistical Analyses. Am. Soc. for Testing
and Materials.

5. Tipper, J. C. 1979. Rarefaction and rareﬁction - the use and abuse of a method in
paleoecology. Paleobiology. 5:423-434.

APPENDIX B

APPENDIX B

STATISTICAL ANALYSIS OF ARDRA DERIVED
RANK-ABUNDAN CE PATTERNS

The following represents an outline for the statistical analysis of ARDRA
derived phylotype-abundance distributions in four steps, and its use is demonstrated for
an analysis of a sequence of four soils of different age (Chapter 3). The tests and their

use are:

1. Chi-Square test. Use the Chi-Square test to show that there is statistical evidence
of a lack of homogeneity across time in the probabilities of SSU rDNA clones being
in the various ranking groups.

2. Kruskal-Wallis test. Use the Kruskal-Wallis test to again provide statistical
evidence of lack of homogeneity. When the null hypothesis is rejected, use Multiple
Comparisons Procedure to pinpoint which of the possible 6 pairings of the 4 time
periods are apparently different.

3. Confidence intervals. Give conﬁdence intervals on the differences between the

true portions in the most dominant phylotypes at different times. Then give similar

107

108
conﬁdence intervals on the differences between the true portions of lower ranking
phylotypes.
4. Trendlines. Create graphics and equations for trendlines ﬁtted for the phylotype-

abundance distribution curves.

To compare rank-abundance patterns from different soil samples we have to
apply tests that are independent of population distributions and associated parameters,
so-called nonparametric tests. This type of test becomes necessary since we cannot yet
make any assumptions about the distribution of the population from which the samples
are drawn. Both the Chi-Square and Kruskal-Wallis tests are non-parametric tests.
They are different approaches for a parallel evaluation. 1 did not use AN OVA here

since it compares averages and ignores the distribution aspects.

1. THE CHI-SQUARE TEST OF HOMOGENEITY
The Chi-Square test is designed to analyze the equality of proportions (Conover,

1980). The Chi-Square test statistic, )8, denotes the difference between observed and

expected frequencies of an event. At x2=0 the observed frequency of events is equal to

the expected one. I applied the Chi-Square test to show that there is statistical evidence
for a lack of homogeneity across different soil samples (1'. e. across time) in the

probabilities of being in the various ranking groups.

109
Contingency table
The observed data are sorted into a contingency table, a two way classiﬁcation

table where the rows represent ordered categories c and the columns symbolize the

different populations k (Table 1) (Johnson et al., 1996).

Table 1: Contingency table to classify and rank different populations in categories

 

 

 

 

 

 

 

 

c categories J
k Populations Category 1: Category 2: Category 0: Row Total
1'
Population 1 01 1 012 01c m
Population 2 02; 022 02c n1
Population 3 031 032 03c n3
Population k 0“ 0),; 0,,c nk
Column total tj t, t; tc Grand total N

 

 

 

 

 

 

k is the number of distinct populations (here: soil communities) from which samples
have been drawn. c is the number of categories into which each sample has been
classiﬁed. Note that the categories are ordered. The indices i and 1' stand for i, rows, and
j, columns. n, is the sample size of the sample from the 1*“ population. Oij is an
observed count in population j that fall into the ith category. The grand total N is deﬁned
as the total number of observations (sum of all m) of all cells in the table. I,- is the sum

of observed counts across k populations within a category

110

Hypothesis
The objective is to determine if the observed proportions in each category are

nearly the same for all populations (Conover, 1980).

Null hypothesis H0: Homogeneity exists across the chronosequence between the four
different clone populations of each clone library. The
probabilities for all populations are the same in each category or
ranking groun H0: 171 =p2 =p3 =p4

Alternative hypothesis H1: Nonhomogeneity across the time periods. Probabilities

for all populations are different in each category or ranking group.

H1:p1¢pz ¢p3 ¢p4

Test Statistic

The objective is to test whether the populations are homogeneous with respect

to cell probabilities (Conover, 1980). In formula [1] the )8 statistic is calculated by

comparing observed counts with expected counts.

12 :2 '1‘) ; [1]

(00' ‘ E‘
E.-,-

The indices i and j stand for i, rows, and j, columns. Oij is an observed count, and Eij is
the expected cell frequency in any table cell if. The expected cell frequency is defrned as

_ (row total for row i)(column total for column j).

E.. , 2
grand total N [ ]

u

 

111

to determine the degrees of freedom for this test statistic we calculate

d.f. = (number of rows - l)(number of columns -1). [3]

The region of rejection for the null hypothesis is deﬁned by a cut off criterion or which
can be found in tables for Chi-Square distribution: x2 2 10,2 [4]

Example

As an example I calculated the test statistic for the chronosequence described in
Chapter 3. The phylotypes were divided into three categories or ranking groups, the
most dominant phylotype in each clone library, the second through the ﬁfth most
abundant phylotype, and ﬁnally all phylotypes in the sixth and higher position. The
subdivision into the three described categories yielded the best comparative results. The
clone libraries constructed for the four soil ages represent the populations in the
contingency table (Table 2). The expected cell frequency Ey' was calculated using

formula [2].

112

Table 2: Contingency table for four soil bacterial communities from soils of different
age. Observed counts are classiﬁed into three ordered categories. The number in
parentheses denote the expected cell frequencies. All data are derived from the 63%
G+C fraction. The data for the two oldest soils (populations k; and k4) are contributed
by three replicates each.

 

 

 

Population: Category 1: Category 2: Category 3:
Soil Age Phylotype l Phylotype Phylotype 6 Row Total
2=>5 plus
k, 200 yrs 10 (20.6) 19 (16.4) 47 (39.0) 76
k2 2,100 yrs 13 (20.9) 18 (16.6) 46 (39.5) 77
k3 20,000 yrs 43 (69.2) 27 (55.0) 185 (130.9) 255
k, 150,000 yrs 109 (64.3) 75 (51.1) 53 (121.6) 237
COIWFO‘“ 175 139 331 645
across trme

 

Following equation [3], the test statistic will have six degrees of ﬁeedom. In a Chi-
Square table we frnd x02 = 12.59 to be the limit at a typical level of signiﬁcance of
p<0.05 and six degrees of ﬁeedom (Johnson et al., 1996). We reject the null hypothesis
H0 if x2212.59. Following equation [1] the test statistic has a value of x2 = 139.2. Since
formula [4] is met we can reject the null hypothesis that there is homogeneity across soil
age for the three ranking groups. Similar calculations lead to a rejection of the null
hypothesis for the 35% G+C fraction, although a Chi-Square test result of 19.4 (p<0.05)

in Chapter 3 indicated much less difference between the four soil ages.

113

 

 

l Most dominant

Dan-Sth
36th

 

 

 

Number of clones

 

200 2,100 20,000 150,000
Soil age (yr)

 

 

Figure 4. Clones per phylotype group for the 63% G+C fractions of a chronosequence
of four soils. The two oldest soils are shown as the mean of three clone library
replicates.

A graphical comparison of determined and expected cell values is shown in
Figure 3 and Figure 4 for the 63% G+C ﬁactions of the chronosequence. The high
diversity contributed by rare members of each clone library (Fig. 3, ranking group 6th-
...) has a strong effect on the Chi-Square value when the dominance pattern becomes
more structured with increasing soil age (Fig. 4). The two oldest soils contribute the

most to the value of Chi-Square, especially their extreme ratios of phylotype diversity

over clone number (Fig. 5).

114

 

40

 

I Most dominant
Uan-Sth

 

 

=61h -...

 

Expected cell counts
[0
O

 

200 2,100 20,000 150,000
Soil age (yr)

 

 

Figure 5. Contribution of expected cell ﬁequencies in the contingency table (Table 2)
to the value of Chi-Square for the 63% G+C fractions of a chronosequence of four soils
(equation [2]).

Since the test statistic had a value of x2 = 139.2 we rejected the null hypothesis Ho of
76212.59.

The Chi-Square test successfully disproved the null hypothesis, and gave statistical
evidence for a lack of homogeneity across soil samples of different age in the

probabilities of being in the three ranking groups.

115

2. THE KRUSKAL-WALLIS TEST
The Kruskal-Wallis test is designed to analyze different, independent

populations of more than two samples to make comparisons between them (Conover,
1980). It is a ﬁmction of the ranks of all observations in the combined sample. After
brieﬂy explaining the test’s theory I will demonstrate how to apply the Kruskal-Wallis
test using actual data from the chronosequence of four soils in Chapter 3.
Assumptions. The following assumptions are made when the Kruskal-Wallis test is
used (Conover, 1980):

1. All samples are random samples ﬁ'om their respective populations

2. There is independence within each sample, as well as mutual

independence among the various samples

3. The measurement scale is ordinal or higher

Hypotheses:

Null hypothesis Ho: All of the population distribution functions are identical. 1. e. , all
four time periods of the chronosequence have rank-abundance
distributions equal for dominance and diversity of the soil
bacterial community.

Alternative hypothesis H]: At least one of the populations tend to yield larger
observations than at least one of the other populations. That is,

at least one of the four time periods of the chronosequence is

116
characterized by a greater degree of dominance and diversity than
at least one of the remaining time periods
Contingency table. Again a contingency table is made (Table 1), and ranks are

computed for each cell. Rj is the average rank in the jth column, and tj is again the column

 

 

 

total,
- t,+1 — t2+l — t3+l — C" (n+1)
=——; R2=t+ ; R_=t+t+ ; ..... RC: ti-l- ; [l]
l 2 l 2 3 l 2 2 j; 2
whereas the sum of ranks on the 1*” row is
a=2m§ m

where c is the number of categories.

Test Statistic T
Rejecting the null hypothesis. If the null hypothesis H0 is true, then the

contributions of the respective ranks E, E, R, in each sample should each be the same

(proportional to sample size n,). Differences between samples change contributions of

E, Ti; E, and increase the test statistic r considerably (Spiegel, 1994). After
calculating the test statistic T the null hypothesis H0 is rejected if T > 121-0, (at the level
of signiﬁcance or). The critical region of a typical level of signiﬁcance 0 = 0.05

corresponds to values of T greater than the 0.95 quantile of a Chi-Square random
variable with k-l degrees of ﬁeedom which can be found in tables for Chi-Square

distribution (Johnson et al., 1996).

117
1 * .2 N(N+l)2
T=—[2—f—) [31

The value of the test variability, 6‘2, is calculated from the equation

1 c — 2 N(N+1)2
S2 = -— t. R. —— 4
~12.“ 4 l ”
Multiple Comparisons Procedure. If H0 is rejected (and only then), the following
multiple comparisons procedure can be employed to determine which pairs of

populations tend to differ (Conover, 1980). The number of possible pairwise

comparisons for k populations is

[SH-21%;) [51

Assume that the ith and j‘h populations are different, if the following inequality is

satisﬁed:

Ri Rj

m "j

”Harm/74— [6]

 

 

where

A-s2(—~-I-T)(i+9
N-k m n;

Riand R,- are the rank sums of the two samples, (1.“); is the (1- a/2) quantile of the (-
distribution with N-k degrees of freedom (see tables for t distribution (Johnson et al.,

1996)).

 

118

This procedure is repeated for all pairs of populations (see Table 4 below as an example
for k=4 populations).
Example

As an example we calculate the test statistic for the chronosequence described in
Chapter 3, four different soil communities (populations k=4), each representing a
different soil age; 72,, n2, n3, 724. are the number of clones investigated for each soil. All
data are derived from the 63% G+C fraction. Again, as for the Chi-Square test above, all
observations are pooled in a contingency table, and the phylotypes were divided into
three categories or ranking groups. The most dominant phylotype in each clone library
is treated as another rank, R, , the second to ﬁfth most dominant clones are treated as
the same rank E , and ﬁnally all of the remaining sixth and higher clones as grouped as a

third rank, 'Rj.

 

The Kruskal-Wallis Test Step by Step

1. Compute average ranks for the different classiﬁcation levels. Then compute
row ranks using both: (a) average ranks from step 1,

and (b) all observed counts

2. Compare the error variance S2

3. Compute the Kruskal-Wallis test statistic T

4. Determine, if T lies within the ‘tail’ of x2

5. Reject homogeneity between populations (soil samples)

6. Use Multiple Comparisons Procedure to compare populations pairwise.

 

 

 

 

119

1. Computing the ranks using a contingency table (Table 1).
First, compute the average ranks E, R; R: for each of the respective ranking groups

using equation [1].

— 175+] — 139+] 331+l

 

 

 

=158; E=88+158+ =412;

Second, compute the rank totals R; s for all rows using equation [2].

R1 = 10(88)+19(158)+ 47(412) = 23,246; same calculation for R2 through R,

2. Determining the test variability S2. The total number of clones in all samples is
N=645. The number of categories is c=3. The value of S2 is calculated using equation

[4]: 52:2898893

3. Rejecting the null hypothesis. Using the result for test variability S2 in equation
[3] the value of the test statistic T is determined as T=125.4. For k-1=3 degrees of
freedom, the quantile found in the Chi-Square distribution table for the common level of

signiﬁcance 0r=0.05 is 7.81. Since the test statistic T is of higher value than 7.81, the

null hypothesis H0 of identical populations and hence, identical population
distributions, is rejected. This allows us to continue with the Multiple Comparisons

Procedure.

4. Multiple Comparisons Procedure. Equation [6] denotes the inequality that has to

be satisﬁed if two populations i and j are called signiﬁcantly different. Following the

120

equation N-k the test statistic will have 641 degrees of freedom. From the tables for t

distribution, the standard variable mm, quantile with N-k =64] degrees of ﬁeedom is ’1-

4
(m) =1.96 for a typical a=0.05. According to equation [5] we can make (2)=6 different

pairwise comparisons of soil samples of different age (populations) in the group of four.
R, and 16,- are the rank sums of two samples each in a pairwise comparison (Table 3).

The value of constant b is obtained using equations [3] and [4]

\ﬁ; =\/(S2NN;;—7:)=153.15.

 

Table 3. Results for the Multiple Comparisons Procedure of four different soil
bacterial communities.

 

 

 

Rr' R" l l
_._ I tr-(a/2)‘/E\/—-—

Populations n_ n}, Statement‘I

I

     

1 and 2 10.79 48.53 False
1 and 3 19.34 39.23 False
1 and 4 144.33 39.57 True
2 and 3 30.13 39.03 False
2 and 4 133.54 39.37 True
3 and 4 163.67 27.08 True

 

" That the compared populations have a different distribution

The Kruskal-Wallis test will indicate if there is a statistically signiﬁcant

difference between any two of the means. In half of the cases the second column

121

exceeds the third column. The Multiple Comparisons Procedure shows that the rank-
abundance distribution of the Kohala clone library at 63% G+C is signiﬁcantly different
from the rank-abundance distributions of all three younger soils (p<0.05). There is,
however, no signiﬁcant difference amongst the three younger soils. The second colmrrn
exceeds the third column in all of the cases for the 35% G+C ﬁ'actions of the
chronosequence. The Multiple Comparisons Procedure shows that the rank-abundance
distribution of the four clone libraries ﬁom the chronosequence 35% G+C ﬁaction
shows no signiﬁcant difference amongst the four soils (p<0.05). However, at a coarser
level of signiﬁcance of p<0.2, in one third of all cases the second column exceeds the
third column. The Multiple Comparisons Procedure reveals that the rank-abundance
distribution of the Ola'a and the Laupahoehoe clone library ﬁ'om the 35% G+C ﬁaction
are signiﬁcantly different from the rank-abundance distributions of the oldest (Kohala)
soil. There is, however, no signiﬁcant difference amongst the three younger soils and

between the youngest and the oldest soil.

 

122

3. CONFIDENCE INTERVALS

To statistically support differences in relative abundance of the most dominant
phylotypes between soils in the chronosequence, I used conﬁdence intervals. A
conﬁdence interval is an estimate. Here we estimate the true probability of a single rrn
clone drawn from the soil and belonging to some dominance rank. The estimate is the
fraction of bacteria in that rank for a speciﬁc sample. Suppose p1 is the unknown
probability of being in the most dominant group for time period 1, and further suppose
p2 is the unknown probability of being in the most dominant group for time period 2. If
a 95% conﬁdence interval estimate is (0.07, 0.19) you are "95% conﬁdent” that the most
dominant group comprised at least 7% more of the population in Time period 1 than the
most dominant group in Time period 2 comprised of the population in the second time
period. If the conﬁdence interval pl-pz contains only positive values we can be
conﬁdent that p, is bigger than p2. The conﬁdence interval shows a lack of homogeneity
if it does not include 0, i. e. there is no statistically signiﬁcant difference.

The Chi-Square- and the Kruskal Wallis-tests test for lack of homogeneity at a
coarser level than conﬁdence intervals. In these tests rank distributions between two
communities are viewed as two separate entities and hence different from each other.
The conﬁdence interval estimates on p1- p2 are more speciﬁc. They allow one to pick a
certain dominance rank, and to estimate how much the contribution of that speciﬁc rank
to the whole population differs across two time groups.

Calculate the conﬁdence interval for p 1- 1);, where p, is the true probability of

being in the most dominant rank group in a certain clone library 1, while p2 is the true

 

123

probability of being in the most dominant rank group in a certain clone library 2. An

approximate 100(1- a)% conﬁdence interval for pl- p; is (Johnson et al., 1996):

 

 

(ﬁr—ﬁ2)iza12\/pi(l—pl)+ [72(1-P2) [1]
n1 n2

for 0<pi<1, and large sample sizes 11, and n; from the two populations that are
compared.

In the example of the chronosequence I assume for the probability multiplier
Zalz = 1.96 which corresponds to a level of signiﬁcance of p <0.05. As an example I
formulated the result for the 63% ﬁ'actions of the chronosequence (Chapter 4): I am
95% conﬁdent that the most dominant group comprised at least 7% more of the clone
library at the oldest site, Kohala, than the most dominant group at the youngest soil,
Thurston. Also, I am 95% conﬁdent that the most dominant group comprised at least
5% more of the clone library at the 20,000 year old Laupahoehoe than the most
dominant group at the youngest soil (Thurston). The conﬁdence intervals of factorial

combinations of other soils did not show signiﬁcant differences.

 

124
4. TRENDLINES
Graphical trendlines are regression curves calculated to ﬁt the phylotype-

abundance distribution curves for the respective soil ages (Fig.6) (Spiegel, 1994). These

trendlines are based on the formula of a geometric curve )7 = ax” , where, in a rank-
abundance distribution, y—values denote phylotype abundance while x-values represent
the rank of the respective phylotype. In comparing curves, the factor or is inﬂuenced by
the number of dominant phylotypes, while factor B is dependent on the frequency of

rare phylotypes, and hence a measure of phylotype diversity. Beta gains inﬂuence with

 

increasing values of x. Trendlines can easily be calculated using the software program

Excel 5.0 (Microsoft Coorporation) (Fig. 6).
.t

b
O

     
 

  
   
   

 

 

 

 

 

 

  

  
   

 

 

Q)

:1.

«>3 y=20,313x '1 .084 $2,033) 150 000 yr
0 9
E I I j 9
0.. 20,000 yr

33.. 20 IIIIII . ...-"......“-

" Y=5.634x '0'5‘9 (R2=O.71)

3 2.100 yr

0) . A

r: I I 7

.9. a I. 200

" r
U 0 .1!'|Ilnl"' IIIII'I II. III . .. .-.". _ _ . . y ‘

 

Phylotypes

Figure 6. Regression trendlines and their equations calculated to ﬁt the rank-abundance
proﬁles for the four soils of the chronosequence. All graphics describe the 63% G+C
fractions for each respective soil site.

125

R2 is the percentage of all variability in the frequencies which is accounted for by the
ﬁtted trend-line in each case (Johnson et al., 1996). Here, R2 is a descriptive statistic,
and I am not assuming any underlying probability distribution. As an example I sorted
the different factors for the four soils of the chronosequence described in Chapter 4

(Table 4).

Table 4. Comparison of factors a and 0 in the regression equation y = ax’3 for
rank-abundance distributions of four soils of different age (63% G+C fraction)

 

 

 

Soil site Soil age (y) Alpha Beta R2

Thurston 200 6.850 -0.575 0.816
Ola’a 2,100 5.634 -0.519 0.707
Laupahoehoe 20,000 2.706 -O.282 0.390
Kohala 150,000 20.313 -1 .084 0.928

 

In this study trendlines are merely used as a graphical approach to show the
trend of diversity versus dominance across age differences, and to express this trend
numercially. There is, however, no statistical merit to trendlines in this study, since we
cannot actually compare them between different soils. Regression analysis treats the
series of phylotypes as a continuum rather than as a series of independent single events

that it is. The strong dependence of residuals (individual deviations from the regression

126

line) is thus not taken into account. Therefore, I did not calculate conﬁdence intervals or
tests of hypotheses based on trendlines. However, trendlines are useful in determining
the similarity of replictates (e. g. 20,000 year series in Table 5). Table 5 summarizes the

results for the trendline equations for all soils investigated.

Table 5. Comparison of factors a and [3 in the geometric regression equation y = our’3
for rank-abundance distributions of all soils investigated (63% G+C fraction)

 

 

Soil Age (y) Vegetationa Alpha Beta R2
200 F 6.850 -0.5746 0.8160
2,100 P 5.634 -0.5189 0.7069
20,000 F- Rep A 2.195 -0.2171 0.3853
20,000 F- Rep B 2.706 -0.2816 0.3899
20,000 F- Rep C 3.654 -0.3742 0.6101
150,000 P- Rep A 9.564 -0.6261 0.7163
150,000 P- Rep B 1.831 -0.7052 0.7053
150,000 F 20.310 -1.0840 0.9278
150,000 F- Rep A 30.505 -1.3242 0. 9548
150,000 F- Rep B 9.564 -1.4311 0. 9443

 

a F stands for forest, P for pasture soils. Rep stands for replicate.

 

127

SUMMARY AND RECOMMENDATIONS

For this summary we use again the example of the Hawaiian chronosequence of
four soils (Chapter 4). The variable in this soil series is soil age. For each soil age we
created rank-abundance proﬁles of phylotypes derived from SSU rDNA clones. The
statistical analysis should determine

(i) if the rank-abundance proﬁles are different from each other [Chi-
Square test], and if so,
(ii) which of all possible pairings of the four soil ages are different
[Kruskal-Wallis test], and ﬁnally,
(iii) amongst those different pairings which rank group of SSU rDNA
clones contributed signiﬁcantly to that difference [Conﬁdence intervals].
This sequence of statistical analyses is appropriate if one wants to compare
independent proﬁles for their signiﬁcant differences. The observed data have to be
arranged in increasing order according to some property such as quality or value (e. g.
rank); samples can have different sizes (e. g. different clone libraries) (Conover, 1980).
The Chi-Square test is less powerful than the Kruskal-Wallis test. In addition to
showing that data collections (e. g. rank-abundance proﬁles) are different the Kruskal -
Wallis test statistic is also a function of the ranks of the observations. However, it
involves more effort than the Chi-Square test. Both, the Chi-Square test as well as the
Kruskal-Wallis test are valid only for large samples (Johnson et al., 1996). Cell

frequencies no smaller than 2 5 are normally required. Replicate analyses of soil from

 

128

the same site are necessary to determine if within site variation between rank-abundance
proﬁles is signiﬁcantly smaller than between site variation.

For a quick descriptive comparison of the four rank-abundance proﬁles without
any statistical assertion trendlines can be made. Trendlines can however be useful to
compare two replicates. The calculations necessary for all tests mentioned are quickly
done with any advanced spreadsheet software (e. g. Excel by Microsoft) or with
common statistical software packages (e. g. Jump by SSPS).

The statistical methods outlined in this Appendix go partly beyond what was
used for this study. Their comprehensiveness should help colleagues who work on
similar questions in ﬁnding the appropriate approach to particular problems with rank-

abundance based community studies.

ACKNOWLEDGMENTS
I thank James Demopolos (Department of Statistics, MSU) for his help in preparing

Appendix B.

REFERENCES

1. Conover, W. J. 1980. Practical Nonparametric Statistics (2nd ed.), pp. 229-234.
John Wiley & Sons, New York, New York, USA.

2. Johnson, R. A. and G. K. Bhattacharyya. 1996. Statistics: Principles and
Methods (3rd ed.). pp. 554-558. John Wiley & Sons, New York, New York, USA.

 

129

3. Spiegel, M. R 1994. Schaum’s Outline Series of Theory and Problems of
Statistics, 2nd ed. McGraw—Hill, Inc., New York, New York, U.S.A.

Recommended reading to expand on the statistical analysis:

Jongman, R. H. G., C. J. F. ter Braak, and O. F. R. van Tongeren (eds.). 1995.
Data Analysis in Community and Landscape Ecology. Cambridge University Press,
Cambridge, England.

BIBLIOGRAPHY

LIST OF REFERENCES

Alexander, M. 1977. Introduction to Soil Microbiology. John Wiley & Sons,
New York, New York, USA.

Amann, R., N. Springer, W. Ludwig, H.-D. Giirtz, and K. H.
Schleifer. 1991. Identification in situ and phylogeny of uncultured bacterial
endosymbionts. Nature 351: 161-164.

Amann, R. I., W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic
identiﬁcation and in situ detection of individual microbial cells without cultivation.
Microbiol. Rev. 59: 143- 169.

Arah, J. R. M., and K. A. Smith. 1985. Anaerobic nricro-environments in
soil and the occurrence of anaerobic bacteria, p. 247-261. In V. Jensen, A.
Kjoller, and L. H. Serensen. (ed.), Microbial Communities in Soil Microbiology.
Elsevier Applied Science Publishers, London, England.

Atlas of Hawaii. 1983. . University of Hawaii Press, Honolulu, Hawaii,
USA.

Betzer, P. R., K. L. Carder, R. A. Duce, J. T. Merill, N. W.
Tindale, M. Uematsu, D. K. Costello, R. W. Young, R. A., Feely,
J. A. Breland, R. E. Bernstein, and A. M. Greco. 1988. Long-range
transport of giant mineral aerosol particles. Nature 336:568-571.

Boivin-Jahns, V., A. Bianchi, R. Ruimy, J. Garcin, S. Daumas,
and R. Christen. 1995. Comparison of phenotypical and molecular methods
for the identiﬁcation of bacterial strains isolated form a deep subsurface
environment. Appl. Environ. Microbiol. 61:3400-3406.

Bomeman, J., P. W. Skroch, K. M. O'Sullivan, J. A. Paulus, N.
G. Rumjanek, J. L. Jansen, J. Nienhus, and E. W. Triplett. 1996.

Molecular microbial diversity of an agricultural soil in Wisconsin. Appl. Environ.
Ecol. 62:1935-1943.

Bomeman, J., and E. W. Triplett. 1997. Molecular microbial diversity in
soils from Eastern Amazonia: evidence for unusual microorganisms and microbial

population shifts associated with deforestation. Appl. Environ. Microbiol.
63:2647-2653.

130

‘1- (I: qum'ouﬂmTR—V

“—3 i h

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

131

Carlquist, S. C. 1980. Hawaii, A Natural History. Paciﬁc Tropical Botanical
Garden, Honolulu, Hawaii, USA.

Clague, D. A., and G. B. Dalrymple. 1981. The Hawaiian-Emperor
volcanic chain: geologic evolution, p.5-54. In R. W. Decker, T. C. Wright, and
P. H. Stauffer (ed.), Volcarrism in Hawaii. United States Geological Survey,
Washington, DC, USA.

Crews, T. E., K. Kitayama, J. H. Fownes, R. H. Riley, D. A.
Herbert, D. Mueller-Dombois, and P. M. Vitousek. 1995. Changes in
soil phosphorous fractions and ecosystem dynamics across a long
chronosequence in Hawaii. Ecology 76: 1407-1424.

Cuddihy, L. W., and C. P. Stone. 1990. Alteration of native Hawaiian
Vegetation: effects of humans, tlrier activities and introductions. Unversity of
Hawaii Press, Honolulu, Hawaii, USA.

D'Aquila, R. T., L. J. Bechtel, J. A. Videler, J. J. Eron, P.
Gorczca, and J. C. Kaplan. 1991. Maximizing sensitivity and specificity of
PCR by preampliﬁcation heating. Nucleic]. Acids Res. 19:3749.

Deharveng, L., and A. Bedos. 1993. Factors inﬂuencing diversity of soil
Collembola in a tropical mountain forest (Doi Inthanon, ‘ Northern Thailand). In
M. G. Paoletti, W. Foissner, D. Coleman (ed.), Soil Biota, Nutrient Cycling, and
Farming Systems. Lewis Publishers, Boca Raton, Florida, USA.

Department of Geography, University of Hawaii. 1983. Atlas of Hawaii,
2nd ed. University of Hawaii Press, Honolulu, Hawaii, USA.

Ellenbroek, F. M., and T. E. Cappenberg. 1991. DNA synthesis and
tritiaited thynridine incorporation by heterotrophic freshwater bacteria in
continuous culture. Appl. Environ. Microbiol. 57: 1675-1682.

Farrelly, V., F. A. Rainey, and E. Stackebrandt. 1995. Effect of
genome size and rrn gene copy number on PCR amplification of 16S rRNA genes
from a mixture of bacterial species. Appl. Environ. Microbiol. 61:2798-2801.

Ferris, M. J., A. L. Ruff-Roberts, E. D. Kopczynski, M. M.
Bateson, and D. M. Ward. 1996. Enrichment culture and rrricroscopy
conceal diverse therm0philic Synechococcus populations in a single hot spring
microbial mat habitat. Appl. Environ. Microbiol. 62: 1045-1050.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

132

Finney, M. 1993. Molecular cloning of PCR products, p. 15.7.1-15.7.6. In F.
M. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A.
Smith, and K. Struhl (ed.), Current Protocols in Molecular Biology, vol. 2.
Greene Publishing Associates and Wiley-Interscience, New York, New York,
USA.

Fox, G. E., E. Stackebrandt, R. B. Hespell, J. Gibson, J.
Maniloff, T. A. Dyer, R. S. Wolfe, W. E. Balch, R. S. Tanner, L.
J. Magrum, L. B. Zablen, R. Blakemore, R. Gupta, L. Bonen, B .
J. Lewis, D. A. Stahl, K. R. Luehrsen, K. N. Chen, and C. R.
Woese. 1980. The phylogeny of prokaryotes. Science 209:457-463.

Fulthorpe, R. R., A. N. Rhodes, and J. M. Tiedje. 1996. Pristine soils
mineralize 3-chlorobenzoate and 2,4-dichlorophenoxyacetate via different
microbial populations. Appl. Envrion. Microbiol. 62:1159-1166.

Giambelluca, T. W., M. A. Nullet, and T. A. Schroeder. 1986.
Rainfall Atlas of Hawaii. Department of Land and Natural Resources, Honolulu,
Hawaii, USA.

Giovannoni, S., T. Britschgi, C. L. Meyer, and K. G. Field. 1990.
Genetic diversity in Sargasso sea bacterioplankton. Nature 345:60—63.

Glasel, J. 1995. Validity of nucleic purities monitored by 260nm/280nm
absorbance ratios. BioTechniques 18:62-63.

Gray, M. W., D. Sankoff, and R. J. Cedergren. 1984. On the
evolutionary descent of organisms and organelles: a global phylogeny based on a
highly conserved structural core in small subunit ribosomal RNA. Nucleic Acids
Res. 12:5837-5852.

Harris, D. 1994. Analyses of DNA extracted from microbial communities, p.
111-118. In K. Ritz, J. Dighton, and K. E. Giller (ed.), Beyond the Biomass.
Compositional and Functional Analysis of Soil Microbial Communities. John
Wiley & Sons, Chichester, England.

Hawaii Scientific Drilling Project Team. 1996. Hawaii scientiﬁc drilling
project: summary of preliminary results. GSA Today 6: 1-8

Herbert, D. A., and J. H. Fownes. 1995. Phosphorus limitation of forest

leaf area and net primary productivity on a weathered tropical soil. Biochemistry
29:223-235.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

133

Hermjakob, H. The rapid identification by fragment length evaluation. [Online]
Available http://bibiserv.techfak.uni-bielefeld.defRIFLE/E.welcome.html,
02/26/1997.

Holben, W. E. 1994. Isolation and puriﬁcation of bacterial DNA from soil, p.
727-751. In R. Weaver et al. (ed.), Methods of Soil Analysis, Part 2.
Microbiological and Biochemical Pr0perties. Soil Science Society of America
Book Series no. 5. Soil Science Society of America, Madison, Wisconsin, USA.

Holben, W. E., and D. Harris. 1995. DNA-based monitoring of total
bacterial community structure in environmental samples. Mol. Ecol. 4:627-631.

Hotchkiss, S. C., and J. O. Juvik. 1993. Pollen record from Kaau Crater,
Oahu, Hawaii: evidence for a dry glacial maximum. Bull. Ecol. Soc. Amer.
74:282.

Hurlbert, S. H. 1971. The nonconcept of species diversity: a critique and
alternative parameters. Ecology 52:577—586.

Huston, M. A. 1994. Biological Diversity. Cambridge University Press,
Cambridge, England.

Innis, M. A., D. H. Gelfand, J. J. Sninsky, and T. J. White (ed.).
1990. PCR Protocols. Academic Press, San Diego, California, USA.

Jackson, M. L., T. W. M. Levelt, J. K. Syers, R. W. Rex, R. N.
Clayton, G. D. Sherman, and G. Uehara. 1971. Geomorphological
relationships of troposherically derived quartz in the soils of the Hawaiian Islands.
Soil Sci.Soc.Am. Proc. 35:515-525.

Jukes, T. H., and C. R. Cantor. 1969. Evolution of protein molecules, p.
21-132. In H. N. Munro (ed.), Mammalian protein metabolism. Academic Press,
New York, New York, USA.

Kim, J., J. R. Cole, E. Torng, and S. Pramanik. 1996. Presented at the
Presented at the Intelligent Systems for Molecular Biology '96. Proceedings of
the Fourth Intemational Conference on Computational Biology, USA.

Kitayama, K., and D. Mueller-Dombois. 1995. Vegetation changes along
gradients of long-term soil development in the Hawaiian montane rainforest zone.
Vegetatio 120:1-20.

41.

42.

43.

45.

46.

47.

48.

49.

50.

134

Kitayama, K. 1996. Soil nitrogen dynamics along a gradient of long-term soil
development in a Hawaiian wet montane rainforest. Plant and Soil 183:253-262.

Komatsoulis, G. A. and M. S. Waterman. 1997. A new computational
method for detection of chimeric 16S rRNA artifacts generated by PCR
ampliﬁcation from mixed bacterial populations. Applied and Environmental
Microbiol. 63:2338-2346.

Kopczynski, E. D., M. M. Bateson, and D. M. Ward. 1994.
Recognition of chimeric small-subunit ribosomal DNAs composed of genes from
uncultivated rrricroorganisms. Appl. Environ. Microbiol. 60:746-748.

Kuske, C. R., S. M. Barns, and J. D. Busch. 1997. Diverse
uncultivated bacterial groups from soils of the arid southwestern United States that
are present in many geographic regions. Appl. Environ. Microbiol. 63:3614-
3621.

Labarca, C., and K. Paigen. 1980. A simple, rapid and sensitive DNA assay
procedure. Anal. Biochem. 102:344-352.

Lee, S.-Y., J. Bollinger, D. Bezdicek, and A. Ogram. 1996.
Estimation of the abundance of an uncultured soil bacterial strain by a competitive
quantitiative PCR method. Appl. Environ. Microbiol. 62:3787-3793.

Levine, S. H. 1976. Competitive interactions in ecosystems. Am. Nat.
110:903-910.

Liesack, W., H. Weyland, and E. Stackebrandt. 1991. Potential risks of
gene ampliﬁcation by PCR as determined by 16S rDNA analysis of a mixed
culture of strictly barophylic bacteria. Microb. Ecol. 21: 191-198.

Liesack, W. and E. Stackebrandt. 1992. Occurrence of novel groups of the
domain Bacteria as revealed by analysis of genetic material isolated from an
Australian terrestrial environment. J. Bacteriol. 174:5072-5078.

Liesack, W., P. H. Janssen, F. A. Rainey, N. L. Ward-Rainey, and
E. Stackebrandt. 1997. Microbial diversity in soil: the need for a combined
approach using molecular and cultivation techniques, p.375-439. In J. D. van
Elsass, J. T. Trevors and E. M. H. Wellington (ed.), Modem Soil Microbiology.
Marcel Dekker, Inc., New York, New York, USA.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

135

Ludwig, W., S. H. Bauer, M. Bauer, I. Held, G. Kirchhof, R.
Schulze, I. Huber, S. Spring, A. Hartmann, and K. H. Schleifer.
1997 . Detection and in situ identiﬁcation of representatives of a widely distributed
new bacterial phylum. FEMS Microbiol. Let. 153:181-190.

Maidak, B. L., G. J. Olsen, N. Larsen, R. Overbeek, M. J.
McCaughey, and and C. R. Woese. 1996. The RDP (Ribosomal Database
Project). Nucleic Acids Res. 24:82-85.

Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning:
a labratory manual. Cold Spring Harbor Labratory, Cold Spring Harbor, New
York, USA.

Mead, D. A., N. K. Pey, C. Herrnstadt, R. A. Marcil, and L. M.
Smith. 1991. A universal method for the direct cloning of PCR ampliﬁed nucleic
acid. Bio/Technology 9:657-663.

Merrill, J. T. 1989. Atmospheric long-range transport to the Paciﬁc Ocean, p.
15-50. In J. P. Riley and R. Chester (ed.), Chemical oceanography, vol. 10.
Academic Press, New York, New York, USA.

Moré, M. I., J. B. Herrick, M. C. Silva, W. C. Ghiorse, and E. L.
Madsen. 1994. Quantitative cell lysis of indigenous microorganisms and rapid
extraction of microbial DNA from sediment. Appl. Microbial. Ecol. 60:1572-
1580.

Moyer, C. L., F. C. Dohbs, and D. M. Karl. 1994. Estimation of
diversity and community structure through restriction fragment length
polymorphism distribution analysis of bacterial l6S rRNA genes from a microbial

mat at an active, hydrothermal vent system, Loihi Seamount, Hawaii. Appl.
Environ. Microbiol. 60:871-879.

Muyzer, G. 1995. Protocols for the DGGE, p. 8-14. In K. Smalla and G.
Muyzer (ed.) EU-Workshop on the Application of DGGE and TGGE in Microbial
Ecology, Braunschweig, Germany. Biologische Bundesanstalt fiir Land- und
Forstwirtschaft, Braunschweig, Germany.

Nannipieri, P., S. Greco, and B. Ceccanti. 1990. Ecological
signiﬁcance of the biological activity in soil, p. 293-356. In J .-M. Bollag and G.
Stotzky (ed.), Soil Biochemistry, vol. 6. Marcel Dekker, Inc., New York, USA.

Nanzyo, M., R. Dahlgren, and S. Shoji. 1993. Chemical characteristics
of volcanic ash soils, p. 145-187. In M. Nanzyo, S. Shoji, and R. Dahlgren
(ed.), Volcanic Ash Soils. Elsevier, Amsterdam, The Netherlands.

61.

62.

63.

65.

66.

67.

68.

69.

70.

136

Niibel, U., B. Engelen, A. Felske, J. Snaidr, A. Wieshuber, R. I.
Amann, W. Ludwig, and H. Backhaus. 1996. Sequence heterogeneities of
genes encoding 16S rRNAs in Paenibacillus polymyxa detected by temperature
gradient gel electrophoresis. J. Bacteriol. 178:5636-5643.

Niisslein, K., and J. M. Tiedje. 1998. Characterization of the dominant and
rare members of a young Hawaiian soil bacterial community with small-subunit
ribosomal DNA ampliﬁed from DNA fractionated on the basis of its guanine and
cytosine composition. Appl. Environ. Microbiol. 64:1283-1289.

Odum, E. P. 1998. Productivity and biodiversity: a two-way relationship. Bull.
Ecol. Soc. Am. 79:125.

Ogram, A., G. S. Sayler, and T. Barkay. 1987. The extraction and
puriﬁcation of microbial DNA from sediments. J. Microbiol. Methods 7:57-66.

Olsen, G. J., D. J. Lane, S. J. Giovannoni, N. R. Pace, and D. A.
Stahl. 1986. Microbial ecology and evolution: a ribosomal RNA approach. Ann.
Rev. Microbiol. 40:337-365.

Olsen, G. J., H. Matsuda, R. Hagstrom, and R. Overbeek. 1994.
fastDNAml: A tool for construction of phylogenetic trees of DNA sequences using
maximum likelihood. Comput. Appl. Biosci. 10:41-48.

Olsen, G. J., C. R. Woese, and R. Overbeek. 1994. The winds of
(evolutionary) change: breathing new life into microbiology. J. Bacteriol. 176:1-

Paul, E. A. and F. E. Clark. 1996. Soil Microbiology and Biochemistry,
2nd ed. Acaderrric Press, San Diego, California, USA.

Peck, T. R., D. Megel, K. Eik, , D. A. Whitney, D. Warncke, R.
C. Munter, J. R. Brown, D. Knudsen, W. C. Dahnke, D. J. Eckert,
D. Beegle, P. E. Fixen, and E. E. Schulte. 1988. Reconurrended
Chemical Soil Test Procedure for the North Central Region. North Dakota
Agricultural Experiment Station, North Dakota State University, Fargo, North
Dakota.

Reiners, W. A., A. F. Bouwman, W. F. J. Parsons, and M. Keller.
1994. Tropical rainforest conversion to pasture: changes in vegetation and soil
prOperties. Ecol. Appl. 4:363-377.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

137

Reysenbach, A.-L., L. J. Giver, G. S. Wickham, and N. R. Pace.
1992. Differential ampliﬁcation of rRNA genes by polymerase chain reaction.
Appl. Environ. Microbiol. 58:3417-3418.

Rheims, H., C. Sproer, F. A. Rainey, and E. Stackebrandt. 1996.
Molecular biological evidence for the occurrence of uncultured members if the
actinomycete line of descent in different environments and geographical locations.
Microbiology 142:2863-2870.

Riley, R. H., and P. M. Vitousek. 1995. Nutrient dynamics and nitrogen
trace gas ﬂux during ecosystem development in montane rain forest. Ecology
76:292—304.

Robinson-Cox, J. F., M. M. Bateson, and D. M. Ward. 1995.
Evaluation of the nearest-neighbor methods for detection of chimeric small-
subunit rRNA sequences. Appl. Environ. Microbiol. 61:1240-1245.

Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. G.
Higuchi, G. T. Horn, K. B. Mullis, and H. A. Ehrlich. 1988. Primer-
directed enzymatic ampliﬁcation of DNA with a therrnostable DNA polymerase.
Science 239:487-491.

Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method
for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425.

Schmidt, T. M., E. F. DeLong, and N.R. Pace. 1991. Analysis of a
marine picoplankton community by 16S rRNA gene cloning and sequencing. J.
Bacteriol. 173:4371-4378.

Schmidt, T. M., B. Pace, and N.R. Pace. 1991. Detection of DNA
contamination in Taq polymerase. BioTechniques 11:176-177.

Simberloff, D. 1978. Use of rarefaction and related methods in ecology, p.
150-165. In K. L. Dickson, J. Cairns, Jr., and R. J. Livingston (ed.), Biological
Data in Water Pollution Assessment: Quantitative and Statistical Analyses.
American Society for Testing and Materials, West Conshohocken, Pennsylvania,
USA.

Snrheim, R., V. L. Torsvik, and J. Gokseyr. 1989. Phenotypical
divergences between populations of soil bacteria isolated on different media.
Microb. Ecol. 17:181-192.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

138

Soil Survey Staff. 1973. Soil survey of the islands of Kauai, Oahu, Maui,
Molokai, and Lanai, state of Hawaii. US. Government Printing Ofﬁce,
Washington, DC, USA.

Spaans, E. J. A., G. A. M. Baltissen, J. Bouma, R, Miedema, A.
L. E. Lansu, D. Schoonderbeek, and W. G. Wielemaker. 1989.
Changes in physical properties of young and old volcanic surface soils in Costa
Rica after clearing of tropical rain forest. Hydrogeol. Processes 3:383-392.

Stackebrandt, E., W. Liesack, and B. M. Goebel. 1993. Bacterial
diversity in a soil sample from a subtropical Australian environment as detenrrined
by 16S rDNA analysis. FASEB J. 7:232—236.

Strunk, O., 0. Gross, B. Reichel, M. May, S. Hermann, N .
Stuckmann, B. Nonhoff, M. Lenke, A. Ginhart, T. Ludwig, A.
Bode, K.-H. Schleifer, and W. Ludwig. 1998. Arb: A software
environment for sequence data. Nucleic Acids Res., in press. Also,
http://www.mikro.biologie.tu-muenchen.de/ARB.

Suzuki, M. T., E. B. Sherr, and B. F. Sherr. 1993. DAPI direct
counting underestimates bacterial abundances and average cell-size compared to
A0 direct counting. Limnol. Oceanogr. 38: 1566-1570.

Suzuki, M. T., and S. J. Giovannoni. 1996. Bias caused by template
annealing in the ampliﬁcation of mixtures of 16S rRNA genes by PCR. Appl.
Environ. Micrbiol. 62:625-630.

Swofford, D. L. 1990. PAUP: Phylogenetic Analysis Using Parsimony,
Version 3.0. Illinois Natural History Survey, Champaign, Illinois, USA.

Tiedje, J. M., J.-Z. Zhou, K. Niisslein, C. L. Moyer, and R. R.
Fulthorpe. 1997. Extent and patterns of soil microbial diversity, p. 35-41. In
M. T. Martins et al. (ed.) Progress in Microbial Ecology. Brazilian Society for
Microbiology, Sao Paulo, Brazil.

Torsvik, V., J. Goksoyr, and F. L. Daae. 1990. High diversity in DNA
of soil bacteria. Appl. Environ. Microbiol. 56:782-787.

Torsvik, V., J. Goksayr, F. L. Daae, R. Serheim, J. Michalsen,
and K. Salte. 1994. Use of DNA analysis to determine the diverstiy of
microbial communities, p.39-48. In K. Ritz, J. Dighton and K. E. Giller (ed.),
Beyond the Biomass. John Wiley & Sons, Chichester, England.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

139

Tsai, Y.-L., and B. H. Olsen. 1991. Rapid method for direct extraction of
DNA from soils and sediments. Appl. Environ. Microbiol. 57: 1070-1074.

Ueda, T., Y. Suga, and T. Matsuguchi. 1995. Molecular phylogenetic
analysis of a soil microbial community in a soybean ﬁeld. Eur. J. Soil Sci.
46:415-421.

Vitousek, P. M. 1982. Nutrient cycling and nutrient use efﬁciency. Am.Nat.
119:553-572.

Vitousek, P. M., and L. R. Walker. 1989. Biological invasion by Myrica
faya in Hawaii: plant demography, nitrogen ﬁxation, and ecosystem effects. Ecol.
Monographs. 59:247-265.

Vitousek, P. M., D. R. Turner, and K. Kitayama. 1995. Foliar
nutrients during long-term soil development in Hawaiian montane rain forest.
Ecology 76:712-720.

Vitousek, P. M., O. A. Chadwick, T. E. Crews, J. H. Fownes, D.
M. Hendricks, and D. Herbert. 1997. Soil and ecosystem develooment
across the Hawaiian Islands. GSA Today, in press.

Wada, K. (ed.). 1989. Allophane and Irnogolite, vol. No.1, Madison,
Wisconsin, USA.

Wang, G. C.-Y., and Y. Wang. 1997. Frequency of formation of chimeric
molecules as a consequence of PCR coampliﬁcation of 16S rRNA genes from
mixed bacterial genomes. Appl. Environ. Microbiol. 63:4645-4650.

Ward, D. M., R. Weller, and M. M. Bateson. 1990. 16S rRNA
sequences reveal numerous uncultured microorganisms in a natural community.
Nature 345:63-65.

Ward, D. M., M. M. Bateson, R. Weller, and A. L. Ruff-Roberts.
1992. Ribosomal RNA analysis of microorganisms as they occur in nature, p.
219-286. In K. C. Marshall (ed.), Advances in Microbial Ecology, vol. 12.
Plenum Press, New York, New York, U.S.A.

Weisblum, B., and E. Haenssler. 1974. Fluorometric properties of the
bisbenzirrridazole derivative Hoechst 33258, a ﬂuorescent probe speciﬁc for AT
concentration in chromosomal DNA. Chromosoma 46:255-260.

102.

103.

104.

105.

106.

107.

108.

109.

140

Weisburg, W. G., J. G. Tully, D. L. Rose, J. P. Petzel, H.
Oyaizu, D. Yang, L. Mandelco, J. Sechrest, T. G. Lawerence, J.
Van Etten, J. Maniloff, and C. R. Woese. 1989. A phylogenetic analysis
of the mycoplasmas: basis for their classiﬁcation. J. Bacteriol. 171:6455-6467.

Weisburg, W. W., S. M. Barns, D. A. Pelletier, and D. J. Lane.
1991. 16S ribosomal DNA ampliﬁcation for phylogenetic study. J. Bacteriol.
173:697-703.

Weller, R., J. W. Weller, and D. M. Ward. 1991. 16S rRNA sequences
of uncultivated hot spring cyanobacterial mat inhabitiants retrieved as randomly
primed cDNA. Appl. Environ. Microbiol. 57:1146-1 151.

Woese, C. R. 1987. Bacterial evolution. Microbiol. Rev. 51:221-271.

Woese, C. R., O. Kandler, and M. L. Wheelis. 1990. Towards a natural
system of organisms: Proposal for the domains Archaea, Bacteria, and Eucarya.
Proc. Natl. Acad. Sci. 87:4576-4579.

Wright, T. C., and R. T. Helz. 1987. Recent advances in Hawaiian
petrology and geochemistry, p.625-640. In R. W. Decker, T. C. Wright, and P.
H. Stauffer (ed.), Volcanism in Hawaii. US. Geological Survey Professional
Paper 1350, United States Government Printing Ofﬁce, Washington, DC, USA.

Zhou, J., M. A. Bruns, and J. M. Tiedje. 1996. Rapid method for the
recovery of DNA from soils of diverse composition. Appl. Environ. Microbiol.
62:316-322.

Zhou, J., M. E. Davey, J. B. Figueras, E. Rivkina, D. Gilichinsky,
and J. M. Tiedje. 1997. Phylogenetic diversity of a bacterial community
determined from Siberian Tundra soil DNA. Microbiology 143:3913-3919.

nICHIcaN stars UNIV. LIBRARIES
llill"llllllllllllllllllllllllllllllllllWHll
31293016885323