dctr final coverpage maresca - mats utc€¦ · concrete is the most widely used building material...

MICROBIAL BIOMARKERS FOR ASR-DAMAGED CONCRETE

Date: February 2018

Julia A. Maresca, PhD, Associate Professor, University of Delaware E. Anders Kiledal, Graduate Research Assistant, University of Delaware

Prepared by: University of Delaware Delaware Center for Transportation Research 301 DuPont Hall Newark, DE 19716 Prepared for: Virginia Center for Transportation Innovation and Research 530 Edgemont Road Charlottesville, VA 22903

FINAL REPORT

1. Report No.

2. Government Accession No. 3. Recipient’s Catalog No.

4. Title and Subtitle MICROBIAL BIOMARKERS FOR ASR-DAMAGED CONCRETE

5. Report Date 4/30/2018

6. Performing Organization Code

7. Author(s) Julia A. Maresca and E. Anders Kiledal

8. Performing Organization Report No.

9. Performing Organization Name and Address University of Delaware Department of Civil and Environmental Eng. 301 DuPont Hall Newark, DE 19716

10. Work Unit No. (TRAIS

11. Contract or Grant No. DTRT13-G-UTC33

12. Sponsoring Agency Name and Address US Department of Transportation Office of the Secretary-Research UTC Program, RDT-30 1200 New Jersey Ave., SE Washington, DC 20590

13. Type of Report and Period Covered Final 7/1/15 –2/28/18

14. Sponsoring Agency Code

15. Supplementary Notes

16. Abstract Concrete is the most widely used building material in the world, and deterioration of concrete infra-structure is likewise a global problem. Current methods for identification of damage rely heavily on hu-man observation, forcing departments of transportation as well as other agencies to replace structures after damage has already occurred. Bio-indicators, primarily bacteria associated with specific risk factors for disease, have long been used to identify risk in food production, water treatment, and recreational waters. Here, we proposed the identification of bacteria that could serve as bio-indicators for early stages of alkali-silica reaction (ASR), a common mechanism contributing to concrete deterioration in the Mid-Atlantic region. Because the chemistry of ASR-affected concrete differs from that of undamaged concrete, we predicted that the microbial population in ASR-affected concrete would also be different. Identification of bioindicators for early detection of ASR could allow state and local departments of transportation to treat ASR, or provide them with more time to budget and plan the replacement of affected structures. This work had two phases: Phase I included proof-of-principle experiments demonstrating that viable bacteria were present in concrete, and that DNA could be extracted from concrete and used to comprehensively identify the bacteria in concrete. Phase II was a 2-year time series analyzing DNA in ASR-reactive and ASR-mitigated concrete samples. The samples weathered outside and were collected every 4-8 weeks over the course of the series. DNA was extracted from all samples and analyzed to monitor changes in the concrete “microbiome” over time. This work demonstrated that viable bacteria inhabit ordinary concrete and that DNA can be easily extracted from small (~5-g) samples. We further showed that these bacteria primarily come from the large aggregate and that some of the bacteria in ASR-affected concrete are different from those in unaffected concrete. Organisms in the Staphylococcus, Aeromicrobium, Chitinophaga, Sediminibacterium, and Xenococcus genera may be useful bio-indicators, and future work will examine these in more detail. This method can potentially be applied to other types of chemical damage affecting concrete.

17. Key Words Concrete, alkali-silica reaction, bio-indicators, bacteria, DNA

18. Distribution Statement No restrictions. This document is available from the National Technical Information Service, Springfield, VA 22161

19. Security Classif. (of this report) Unclassified

20. Security Classif. (of this page) Unclassified

21. No. of Pages XX

22. Price

1

EXECUTIVE SUMMARY

Concrete is the most widely used building material in the world, and deterioration of concrete infrastruc-ture is therefore a global problem. Current methods for identification of damage concrete rely heavily on human observation of visible damage, which is often only apparent after extensive internal damage has occurred. This often gives departments of transportation and other agencies very short time frames for planning, repair, or replacement of affected structures. Bio-indicators, primarily bacteria associated with specific risk factors for disease, have long been used to identify risk in food production, water treatment, and recreational waters. Here, we proposed the identification of bacteria that could serve as bio-indica-tors for early stages of alkali-silica reaction (ASR), a common mechanism contributing to concrete deterio-ration in the Mid-Atlantic region. Because the chemistry of ASR-affected concrete differs from that of un-damaged concrete, we predicted that the microbial population in ASR-affected concrete would also be different. Identification of bioindicators for early detection of ASR could allow state and local depart-ments of transportation to treat ASR, or provide them with more time to budget and plan the replace-ment of affected structures.

This work had two phases: Phase I included proof-of-principle experiments demonstrating that viable bac-teria were present in concrete, and that DNA could be extracted from concrete and used to comprehen-sively identify the bacteria in concrete. Phase II was a 2-year time series analyzing DNA in ASR-reactive and ASR-mitigated concrete samples. The samples weathered outside and were collected every 4-8 weeks over the course of the series. DNA was extracted from all samples and analyzed to monitor changes in the concrete “microbiome” over time. This work demonstrated that viable bacteria inhabit or-dinary concrete and that DNA can be easily extracted from small (~5-g) samples. We further showed that these bacteria primarily come from the large aggregate and that some of the bacteria in ASR-affected concrete are different from those in unaffected concrete. Organisms in the Staphylococcus, Aeromicro-bium, Chitinophaga, Sediminibacterium, and Xenococcus genera may be useful bio-indicators, and future work will examine these in more detail. This method can potentially be applied to other types of chemical damage affecting concrete.

2

FRONT MATTER

Acknowledgments:

This project was funded by the Mid-Atlantic Transportation Sustainability Center – Region 3 University Transportation Center (MATS UTC). Figures 4 and 5 and Table 3 are reproduced with permission from (Maresca et al., 2017).

We gratefully acknowledge colleagues at the Delaware Department of Transportation, especially David Dodd, for providing the materials for the Phase II concrete samples. Dr. Ardeshir Faghri and Dr. Thomas Schumacher at the University of Delaware were instrumental in the early stages of this project. We thank Gary Wenczel, in the UD Structures Laboratory, for his assistance in mixing and pouring the concrete cyl-inders. Dr. Jessica Keffer helped to develop the DNA extraction procedure, validated it, and did the initial sequence analysis. Anders Kiledal did the majority of the analysis presented in Figs. 6-12. Several under-graduate researchers, including Mary Katherine Sutter, Paul Moser, and Alison Treglia, also contributed to this work.

Disclaimer:

“The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the information presented herein. This document is disseminated under the sponsorship of the U.S. Department of Transportation’s University Transportation Centers Program, in the interest of in-formation exchange. The U.S. Government assumes no liability for the contents or use thereof.”

3

EXECUTIVE SUMMARY ............................................................................................................... 1

Front Matter ............................................................................................................................... 2

Figures................................................................................................................................................. 4

Tables .................................................................................................................................................. 4

1. INTRODUCTION ...................................................................................................................... 5

1.1 Problem ......................................................................................................................................... 5

1.2 Approach ....................................................................................................................................... 6

2. METHODOLOGY ...................................................................................................................... 7

2.1 Proof-of-principle experiments ..................................................................................................... 7 2.1.1 Production of concrete cylinders for proof-of-principle experiments ..................................................... 7 2.1.2 Cultivation of viable bacteria from concrete. ......................................................................................... 7 2.1.3 Development of DNA extraction method ............................................................................................... 8 2.1.4 DNA sequence analysis (Phase I) ........................................................................................................... 9

2.2 ASR test series ............................................................................................................................... 9 2.2.1 Preparation of ASR-prone and ASR-mitigated concrete test cylinders .................................................... 9 2.2.2 DNA extraction and sequencing from time series .................................................................................10 2.2.3 DNA sequence analysis and identification of putative bioindicators ......................................................10

3 FINDINGS ............................................................................................................................... 10

3.1 Proof-of-principle experiments ................................................................................................... 10 3.1.1 Culture collection .................................................................................................................................11 3.1.2 Proof-of-principle: DNA analysis of test sample ....................................................................................12

3.2 ASR Bioindicators ........................................................................................................................ 15 3.2.1 Overview of sequencing results ............................................................................................................15 3.2.2 Sources of bacteria in concrete ............................................................................................................17 3.2.3 Changes in bacterial assemblages over time .........................................................................................17 3.2.4 Differences between ASR-reactive and ASR-mitigated concrete ............................................................19

4 CONCLUSIONS........................................................................................................................ 21

4.1 General conclusions .................................................................................................................... 21

4.2 Caveats ........................................................................................................................................ 22

4.3 Proposed bioindicators and Future work .................................................................................... 23

5 References ............................................................................................................................. 24

6 Appendices ........................................................................................................................ 27

Appendix 1: Invited seminars presenting the funded work. .............................................................. 27

Appendix 2: Conference presentations of the funded work .............................................................. 28

Appendix 3. Publication: Maresca et al., Materials and Structures 2017 .......................................... 29

4

FIGURES

Figure 1. Visible map cracking in ASR-affected concrete.. ................................................................................ 6

Figure 2. Cultivation strategy.. ............................................................................................................................ 7

Figure 3. DNA extracted from concrete.. ........................................................................................................... 8

Figure 4. Microbes identified in proof-of-principle sequence analysis.. ........................................................ 13

Figure 5. Rarefaction curves.. ........................................................................................................................... 14

Figure 6. Experimental plan for identifying potential bio-indicators for ASR in concrete.. ........................... 15

Figure 7. Comparison of bacteria found in ASR-reactive and ASR-mitigated concrete. ................................ 18

Figure 8. Sources of bacteria in concrete......................................................................................................... 19

Figure 9. Changes in phylum abundance over time. ....................................................................................... 19

Figure 10. Changes over time in the differentially abundant genera identified in Figure 6. ........................ 20

Figure 11. Changes over time in selected differentially abundant genera. ................................................... 21

Figure 12. E. coli on concrete surface. ............................................................................................................. 23

TABLES

Table 1. DelDot Class B mix design.. ................................................................................................................... 7

Table 2. Mix designs for ASR-reactive (class A-503) and ASR-mitigated concrete (class A-3-50). ................ 10

Table 3. Bacteria isolated from concrete. ........................................................................................................ 11

Table 4. Sequence data from phase II experiment. ......................................................................................... 17

5

F I NAL REPORT : MI CR OBI AL B I OMAR KERS FOR ASR -DAMAGED C ONCRET E

1. INTRODUCTION

1.1 PROBLEM

Concrete is the most widely used building material in the world (Chen and Liew, 2003). Although concrete can be long-lasting – the Romans used concrete to build bridge piers, jetties, and even the dome of the Pantheon nearly 2,000 years ago (Delatte, 2001) – most concrete structures are built with a finite pre-dicted service lifetime of ~100 years (Wegen et al., 2012). As these structures age, detection of damage or potential damage becomes important for both planned use and planning for replacement or repair. Currently, structural health is assessed using primarily visual methods, which can be subjective (Arndt et al., 2011). For this reason, there is a serious need for simple, minimally-invasive diagnostic tests that can identify damage prior to visible deterioration.

In food, health, and water infrastructure, specific kinds of bacteria or other microbes associated with spe-cific problems are used as “bioindicators” to identify risks of disease or contamination that would other-wise be undetectable (Amini and Kraatz, 2014; Marine et al., 2015). Based on the routine use of bioindi-cators in other systems, we believe that they may also be applicable to evaluation of concrete infrastruc-ture as well. Deterioration mechanisms such as alkali-silica reaction (ASR), carbonation, or corrosion of reinforcing steel cause chemical changes in the concrete that alter the environment inhabited by mi-crobes. These chemical changes could potentially lead to observable changes in the assemblage of mi-crobes in concrete, making it more hospitable to some microbes and less hospitable to others. For this reason, the presence of particular species of bacteria may indicate the physical and chemical conditions in concrete that could lead to structural damage. However, bacteria known to be associated with con-crete damage are associated with biofilms on concrete surfaces, not internal to concrete structures (reviewed in Wei et al., 2010). Environmental conditions on the surface of concrete can be quite different from the internal conditions, and different bacteria would be expected to live inside the concrete.

Therefore, to assess the potential for use of bioindicators in structural evaluation, we must first confirm that viable bacteria are present in concrete, then develop a method for identifying these bacteria. Some bioindicator tests, such as enumeration of fecal coliforms in recreational or drinking waters, require culti-vation of bacteria in the laboratory. Because the bio-indicator microbes in these tests may be patchily dis-tributed, present in very low concentrations, or even unrelated to the risk assessed, cultivation-based as-says are not currently regarded as the best practice (Harwood et al., 2014). More recently, methods have

6

been developed that detect genes unique to bioindicator species in the environment or sample being evaluated (Harwood et al., 2014). These methods are both more sensitive and more robust than culture-based tests, but require extraction of high-quality DNA from the sample. The goals of this research were thus to (1) perform proof-of-principle experiments to show that bacteria in and on concrete could be de-tected and analyzed, then(2) investigate the potential for identification and application of bioindicators to detect ASR damage in concrete.

ASR is a chemical reaction between the alkaline cementitious materials and reactive silica present in cer-tain aggregates, and is a common mechanism contributing to concrete deterioration in the Mid-Atlantic (Touma et al., 2001). When there are available cations in the concrete pore solution, reactive silica in the aggregate, and sufficient moisture, a gel-like material forms in the concrete, absorbs water and expands, exerting tensile forces within the concrete matrix and causing cracking (Fig. 1). External water easily pen-etrates the cracked concrete, exacerbating ASR and increasing the potential for other kinds of damage, further deteriorating the concrete. Because the chemistry of ASR-affected concrete differs from that of

undamaged concrete, we predicted that its microbial popu-lation would also be different. Identification of bioindicators for early detection of ASR could allow state and local depart-ments of transportation to treat ASR, or provide them with more time to budget and plan the replacement of affected structures.

Here, we demonstrate that viable bacteria inhabit ordinary concrete and that DNA can be easily extracted from small (~5-g) samples. We further show that these bacteria primar-ily come from the large aggregate, that some of the bacteria in ASR-affected concrete are different from those in unaf-fected concrete, and that Staphylococcus, Aeromicrobium, Chitinophaga, Sediminibacterium, and Xenococcus are poten-tially useful bio-indicators.

1.2 APPROACH

In order to demonstrate that bacteria in concrete are alive and can be analyzed, we first cultivated some of the viable bacteria in concrete (Fig. 2), then developed a method for extraction of DNA from concrete. After analysis of this DNA confirmed that this approach would be feasible, we prepared 15 small test cyl-inders using concrete materials from the Delaware Department of Transportation (DelDOT) known to be highly susceptible to ASR, and a parallel series of 15 test cylinders that incorporated DelDOT’s standard mitigation against ASR. All cylinders were placed outside on a roof to weather, and one cylinder from each series was removed from the roof every 6-8 weeks and archived. DNA was extracted from three sub-samples of each cylinder, bacterial 16S genes were amplified and sequenced using the Earth Microbiome Project protocol (Gilbert et al., 2014; Thompson et al., 2017), and sequences were analyzed using Qiime2 software (https://qiime2.org/).

FIGURE 1. VISIBLE MAP CRACKING IN ASR-AFFECTED CONCRETE. The concrete slab on the left displays surface cracks typical of ASR. The concrete slab on the right is unaf-fected.

7

2. METHODOLOGY

2.1 PROOF-OF-PRINCIPLE EXPERIMENTS

2.1.1 Production of concrete cylinders for proof-of-principle experiments

To determine whether ordinary concrete is colonized by microbes, concrete cylinders (152 × 305 mm) were cast in December 2011 using a Delaware Department of Transportation (DelDOT) Class B mix design (Table 1), which is typically used for concrete pavements. Crushed coarse aggregates had a maximum size of 19 mm.

After casting, the cylinders were cured in capped plas-tic forms for 2 weeks, then uncapped and kept in a storage room at room temperature until February 2012, when two cylinders were placed on the roof of a University building in Newark, Delaware, where they were exposed to weather. After 12 months of expo-sure, samples for cultivation and DNA analysis were collected.

2.1.2 Cultivation of viable bacteria from concrete.

Prior to sampling, cylinders were moved from the rooftop to a pre-sterilized sampling area. All tools were sterilized with 95 % ethanol. The top surface of one cylinder was struck using a 3-in. masonry chisel and hammer until surface chips or larger chunks spalled off. Individual chips were transferred to sterile buffer (10 mM Tris, 1 mM EDTA) in 50-mL centrifuge tubes. The samples were vortexed for 10 s and 25 μL of the buffer was spread on Petri dishes filled with solid CM-A medium, a minimal, slightly alkaline medium de-veloped for this study. CM-A contains (per liter), 1 g NaHCO3, 1.44 g ammonium acetate, 0.12 g

Characteristic Specification Design compressive

strength, fc’ 20 MPa (2900 psi)

Design cement con-tent (minimum) 334 kg/m3 (560 lb/yd3)

Design water-to-ce-ment ratio, w/c 0.45

Required air content 4–7 % Required slump 50–100 mm (2–4 in.)

Required admixtures (ASTM 2013) A, D, E, F, G

TABLE 1. DELDOT CLASS B MIX DESIGN. Specifica-tions for the concrete used for proof-of-principle experiments were obtained from the Delaware De-partment of Transportation.

FIGURE 2. CULTIVATION STRATEGY. Concrete chips were dropped into sterile saline solution and mixed, then the solution was spread onto growth medium and incubated un-til colonies appeared. Individual colonies were then restreaked on fresh medium until only one kind of bacteria was present.

8

MgSO4·7H2O, 0.14 g KH2PO4, 0.17 g K2HPO4, 1 mL each Wolfe’s vitamin and trace minerals solutions (Wolin et al., 1963), and 11 mL phosphate buffer (per liter, 62.7 g KH2PO4 and 1.27 g K2HPO4, pH 8.5). The pH of the medium was adjusted to 8.0. The Petri dishes were incubated at room temperature (~23°C) for 2 weeks. Individual colonies were restreaked onto fresh CM-A media until only one colony type was visi-ble. Isolation of bacterial strains was verified by microscopy and 16S rRNA sequencing. Once isolated, strains were archived at −80 °C in 10 % glycerol.

2.1.3 Development of DNA extraction method

Subsamples of the concrete cylinders were obtained first by breaking off pieces of the concrete as de-scribed in section 2.2, then grinding ~5 g samples to powder in a mortar and pestle.

To facilitate comparison of the sequence data to data obtained from other environments, we followed the recommendations of the Earth Microbiome Project to the degree possible (http://www.earthmicrobiome.org/). For this reason, DNA extraction from concrete was initially at-tempted using the MoBio PowerSoil and PowerMax DNA extraction kits (MoBio, Carlsbad, CA, catalog nos. 12888 and 12988, respectively), but no genomic DNA was obtained. We then attempted a method developed for DNA extraction from soil, with a high-salt, high-temperature lysis step (Zhou et al., 1996), but again no DNA was obtained.

Concrete is dry, hard, and high in divalent cations such as Ca2+. Other materials with similar characteristics include limestone, mineral-rich sediments, and dehydrated bone. A method used for extraction of DNA from endolithic communities was also attempted (Hugenholtz et al., 1998), but no DNA was obtained us-ing this method. Another method developed for extraction of nucleic acids from high-silica sediments, with an alkaline incubation at high temperature to dissolve the silicate minerals (Kouduka et al., 2012), was also attempted without success.

Subsequently, methods developed for extraction of DNA from ancient bone were modified for application to concrete samples (Cattaneo et al., 1995; Loreille et al., 2007). In these methods, the sample is washed with EDTA (ethylenediaminetetraacetic acid) prior to extraction, to remove some of the divalent cations, then additional salts are precipitated using a sodium acetate solution. In the method we developed for

concrete, each 5-g sample of crushed concrete was washed with 20 mL of 0.5 M EDTA. Be-cause so many of the bacteria we isolated were Gram-positive bacteria with membranes that can be difficult to disrupt, we also in-cluded two enzymatic lysis steps. The EDTA-washed powder was suspended in 15 mL of ly-sis buffer (40 mM EDTA, 50 mM Tris pH 8.3, 0.73 M sucrose) with 15 mg of lysozyme per gram of concrete powder and incubated at 37 °C with gentle agitation for 30 min to break open bacterial cell walls. Next, 100 µL of pro-teinase K (20 mg mL−1) and 1.5 mL of 20 % so-dium dodecyl sulfate (SDS) were added and the solution was incubated at 56 °C for 2.5–3 h

FIGURE 3. DNA EXTRACTED FROM CONCRETE. (A) Ge-nomic DNA extracted from soil (lane 1) and concrete (lanes 2-6). The size standard is in lane “M”. (B) Lanes 1-3: V3-V4 region of the 16s rRNA gene PCR-amplified from DNA extracted from concrete. Lane M: size standard. Empty lanes are negative control reactions.

9

with gentle agitation to break any remaining cell walls. Membranes, proteins and cell debris were sepa-rated from DNA and other water-soluble material by extraction with 20 mL chloroform. To remove water-soluble non-nucleic acid material, 1 mL of 1.95 M sodium acetate was added to the aqueous phase of each sample after extraction and mixed vigorously for 30 s. The sample was then centrifuged and the aqueous phase was transferred to a fresh centrifuge tube and re-extracted with 0.8 volume of chloro-form. DNA was precipitated with 1 volume of isopropanol and 0.1 volume 3 M sodium acetate. The DNA pellet was washed with 70 % ethanol, and genomic DNA was then re-suspended in 25 µL of sterile water and stored at −20 °C. This method has been published in Materials and Structures (Maresca et al., 2017).

Using this protocol, we consistently obtain high-molecular-weight genomic DNA from concrete (Fig. 3A). This DNA is pure enough for downstream processing by polymerase chain reaction (PCR) amplification and sequencing (Fig. 3B). To confirm that biomass is not lost during the EDTA wash, DNA was extracted from the EDTA supernatant after the wash step. No genomic DNA was observed in this fraction, confirm-ing that the additional wash steps do not result in appreciable loss of biomass or DNA from the samples.

2.1.4 DNA sequence analysis (Phase I)

The extracted genomic DNA was sent to Research and Testing Laboratories (Lubbock, TX) for 16S rRNA gene amplification and sequencing. The 16S gene in bacteria is used for species identification. For general analysis of all organisms in an environment, primers that amplify a region of this gene that is found in all species are used, then the genes are sequenced to identify the organisms present. For more specific anal-ysis such as presence/absence of bioindicator species, different primer sets unique to the organisms of interest can be designed. Here, general primers for amplification of the V3–V4 region of the gene were used (515F and 806R (Caporaso et al., 2012)). Sequencing was performed on an Illumina MiSeq instru-ment; both amplification and sequencing were done by Research and Testing Laboratories (Lubbock, TX).

Paired raw reads were joined, quality-filtered and assigned to taxonomic groups using Qiime1 pipelines (Kuczynski et al., 2012; Aronesty, 2013). Operational taxonomic units (OTUs, analogous to species) were defined as gene clusters sharing 97 % sequence identity. Alpha and beta diversity parameters were calcu-lated using standard Qiime pipelines (Kuczynski et al., 2012).

2.2 ASR TEST SERIES

2.2.1 Preparation of ASR-prone and ASR-mitigated concrete test cylinders

Materials for ASR-reactive concrete were obtained from the Delaware Department of Transportation (DelDOT). Two parallel batches of concrete were mixed in the UD Structures Laboratory. The ASR-reactive mix, with no mitigation to prevent ASR, used the mix design for DelDOT Class A-503 concrete (Table 2). The ASR-mitigated mix, which used the same materials but replaced half of the cement with slag, used the mix design for DelDot Class A-3-50 (Table 2). The batches were initially mixed according to the specifi-cations, but some additional water was added to improve workability prior to pouring.

10

Concrete was poured into 4”d x 8”h cylindrical plastic forms, which were agitated to remove air pockets, then capped and placed in a storage room to cure for 2 weeks. After curing, cylinders were removed from

the forms and placed on the roof of Colburn Laboratory on the Uni-versity of Delaware campus, where they would be exposed to wind, sun, and precipitation, but protected from foot and car traf-fic. Two cylinders (one from each series) were archived immediately (May 2, 2013) without being placed on the roof. After this, one cylinder from each series was ar-chived approximately every 4-8 weeks until February 17, 2015.

2.2.2 DNA extraction and sequencing from time series

Subsamples (5 g) were collected from 3 different locations on each cylinder with a drill core (3/4” inch bit) sterilized with ethanol. Each subsample was then crushed to powder in a rock grinder and stored at -80°C until processing. After crushing, DNA was extracted as described in section 2.1.3.

DNA was also extracted from the materials used in the concrete mixes. The same procedure was used to extract DNA from large aggregate, fine aggregate, Portland cement, and slag. To identify any contami-nants that might be introduced during either the extraction process or the amplification and sequencing processes, glass beads (5 g) were sterilized by bleaching, UV irradiation, and autoclaving, then processed as the concrete and materials were.

2.2.3 DNA sequence analysis and identification of putative bioindicators

The V3-V4 region of the 16S genes in all samples was amplified using primers 357F and 806R at the UD Sequencing and Genotyping Center. Samples were then sequenced using an Illumina MiSeq system, with paired 301-base-pair reads. Read pairs were quality-filtered and joined, then grouped into amplicon se-quence variants (ASVs, equivalent to species), and assigned to specific taxonomic groups using standard parameters in Qiime2 (https://qiime2.org/).

3 FINDINGS

3.1 PHASE I: PROOF-OF-PRINCIPLE EXPERIMENTS

ASR-reactive DelDOT Class A-503

Non-reactive (mitigated) DelDot Class A-3-50

Component* Per yd3 Batch Per yd3 Batch Cement 799.0 44.4 353.0 33.9

Slag 0.0 0.0 353.0 33.9 Silica Fume 0.0 0.0 0.0 0.0

Water 267.7 14.9 282.0 27.0 Air 0.0 0.0 0.0 0.0

Fine Aggregates 1070.0 59.4 1064.0 102.1 Coarse Aggregates 1881.0 104.5 1870.0 179.4

Admix 1 0.0 0.0 0.0 0.0 Admix 2 0.0 0.0 0.0 0.0

Total weight 4017.7 223.2 3922.0 217.9

w/c Ratio 0.335 0.34 0.399 0.40 TABLE 2. Mix designs for ASR-reactive (class A-503) and ASR-mitigated concrete (class A-3-50). Mix designs and materials were obtained from DelDOT.

*All units are pounds.

11

To demonstrate that bacteria in concrete are alive and can be analyzed, we first cultivated some of the bacteria in concrete (Fig. 2), then developed a method for extraction of DNA from concrete. Analysis of this DNA confirmed that a DNA-based approach to analysis of bacterial assemblages in and on concrete would be feasible.

3.1.1 Culture collection

Fifteen bacterial strains were isolated from the surface and inside of the concrete samples. Greater diver-sity was initially observed on solid media spread with samples from surfaces than interior samples. Indi-vidual colonies were transferred to fresh solid medium until the culture contained only one kind of organ-ism, as assessed by microscopy and 16S rRNA amplification and sequencing.

Nearly all of the bacteria isolated from the concrete test samples are Actinobacteria, primarily Arthrobac-ter spp. (Table 3). One representative each of Alphaproteobacteria (a Paracoccus species) and Gamma-proteobacteria (Pseudoalteromonas) were also isolated. Most of these strains are brightly colored due to the synthesis of carotenoid pigments (data not shown), and grow readily on both solid and liquid medium using acetate as the carbon and energy source. Relatives of several of these species were isolated from environments that are similar to concrete in salinity or alkalinity. For example, Kocuria species have been found in marine sediments (Kim et al., 2004) and salt evaporation works (Sarafin et al., 2014). Dietzia natronolimnaea strains have been isolated from moderately saline alkaline lakes (Duckworth et al., 1998), and Nocardioides lentus has been isolated from alkaline soils (Yoon et al., 2006).

TABLE 3. BACTERIA ISOLATED FROM CONCRETE. All strains were identified by amplification of the 16s rRNA gene with primers 8F and 1492R and Sanger sequencing of the products. Strains designated “CL” were isolated from top or side surfaces of concrete; strains designated “IN” were isolated from internal samples. They were com-pared to GenBank, the non-redundant gene database maintained at the National Library of Medicine (ncbi.nlm.nih.gov), and the closest cultivated relative of each of our isolates was identified in this database. This table is reprinted with permission from Maresca, Moser and Schumacher, 2017.

Strain des-ignation

Phylum Closest relative in GenBank Colony color Cell Shape

IN-02CB Alphaproteobacteria Paracoccus marinus KKL-B9 Green-Yellow IN-02 Gammaproteobacteria Pseudoalteromonas sp. BSi20669 Orange short rod IN-01 Actinobacteria Arthrobacter sp. 32c White short rod

IN-02C Actinobacteria Kocuria marina Green-Yellow coccus IN-03 Actinobacteria Dietzia cercidiphylli strain X10 Orange rod IN-04 Actinobacteria Rhodococcus cercidiphylli strain

EB37 Orange-Yel-low

long rod

IN-06 Actinobacteria Arthrobacter sp. clone EK_Ca790 Pink rod IN-06B Actinobacteria Kocuria marina Yellow-Green coccus

(pairs) IN-07 Actinobacteria Arthrobacter agilis strain DSM

20550 Pink coccus

IN-08 Actinobacteria Nocardioides lentus strain KSL-19 Burnt Orange IN-09 Actinobacteria Dietzia natronolimnaea strain

NF047 Orange-Red rod

CL-01 Actinobacteria Dietzia natronolimnaea strain NF047

Orange short rod

CL-02 Actinobacteria Uncultured Arthrobacter sp. clone Acro223

Yellow-Green coccus

CL-04 Actinobacteria Arthrobacter sp. 5139 White coccus (pairs)

CL-05 Actinobacteria Rhodococcus erythropolis strain zzx26

coccus

12

Isolation of these bacteria demonstrated that there are bacteria in and on concrete, and that at least some of these bacteria are alive. One concern, prior to this analysis, had been that any bacteria that were incorporated into concrete from the materials would be killed either by the extreme conditions in the concrete (at >12, the pH in concrete is nearly the same as the pH of bleach, which is used for disinfection) or by the high temperature during curing. DNA is, in fact, quite stable, and it is possible that DNA found in concrete could have been “relic DNA” (Carini et al., 2016). This is still a possibility. However, the presence of viable bacteria in concrete demonstrates that at least some of the bacteria in and on concrete are not just alive but may respond to changing conditions.

3.1.2 Proof-of-principle: DNA analysis of test sample

We sequenced only two samples (DNA extracted from a surface subsample, and DNA extracted from an interior subsample) in this proof-of-principle experiment, and obtained ~25,000 DNA read pairs from DNA extracted from the top of the concrete cylinder, and ~59,000 read pairs from the inside sample. Since each 16S gene was sequenced from both ends, this number of reads meant that approximately 24,000 16S genes from the top sample and 49,000 from the inside sample were completely sequenced. More than 250 bacterial genera were identified in these sequence data sets. The most abundant groups in both the interior and surface samples were Actinobacteria and Proteobacteria (Fig. 4). The Micrococcacea, a subset of Actinobacteria that includes the Arthrobacter spp. and Nocardioidaceae that we isolated, were abundant in the interior sample. These organisms, though they do not form spores, are capable of form-ing dormant cyst-like cell types when conditions are unfavorable to growth (Demkina et al., 2000; Young et al., 2010; Shimkets, 2013). Approximately 2.5 % of the sequences in the interior sample belong to the Firmicutes (Bacillus and Clostridium species), which do form spores; in total, Gram-positive bacteria (Ac-tinobacteria and Firmicutes) comprise more than 65 % of the species identified in the interior sample (Fig. 4A). Of the Proteobacteria in the interior sample, the most abundant were Alphaproteobacteria, specifi-cally Rhizobia and Caulobacteria, which are both common soil bacteria. Nearly 6 % of the sequences in this sample could not be assigned to specific taxa, likely because they were equally similar to multiple types of bacteria.

In contrast, the microbial assemblage on the surface of the cylinder was only ~33% Actinobacteria (Fig. 4B). In this sample, the Alphaproteobacteria and Betaproteobacteria species were much greater fractions of the population (Fig. 5B). Unsurprisingly, the surface sample also had more sequences derived from photosynthetic Cyanobacteria (blue-green algae) and chloroplasts (plant-associated organelles that likely came from pollen in these samples) than the interior sample. Spore-forming Firmicutes were an even smaller fraction of the surface sample than in the interior sample. Deltaproteobacteria related to Bdel-lovibrio and Myxococcus, predatory species that consume other bacteria, were observed only in the top sample (Fig. 4B).

The DNA sequence analysis demonstrated that the microbial assemblages had much greater diversity than observed in the culture collection. This discrepancy between cultivation-dependent and cultivation-independent analyses of the same environment is common, since we know very little about the nutri-tional requirements of environmental bacteria (Epstein, 2013; Giannantonio, Kurth, Kurtis, & Sobecky,

13

FIGURE 4. MICROBES IDENTIFIED IN PROOF-OF-PRINCIPLE SEQUENCE ANALYSIS. (A) Microbes found inside concrete sample, identified by 16s rRNA gene sequencing. (B) Microbes found on the surface of the concrete sample, identified by 16s rRNA gene sequencing. In both panels, colors indicate phylum-level identification, and the names are family-level identification. Asterisks (*) indicate that organisms from this group were suc-cessfully cultivated in our laboratory (see Table 3). This figure is reprinted with permission from Maresca, Moser and Schumacher, 2017.

Micrococcaceae

Micrococcaceae (Arthrobacter)

Nocardiodaceae

Rhodobacteraceae (Paracoccus)

Other/unclassified

Oxalobacteraceae

Nocardiodaceae (Rhodococcus)Comamonadaceae

Sphingomonadaceae (Sphingomonas)

Sphingomonadaceae (Kaistobacter)

Micrococcaceae

Bacillaceae (Bacillus)

Methylobacteriaceae (Methylobacterium)Acetobacteriaceae

SporichthyaceaeRhodospirillaceae (Skermanella)

Propionibacteraceae (Propionibacterium)Sphingomonadaceae

EnterobacteriaceaeErythrobacteriaceae

Nocardiodaceae (Nocardioides)Chloroplast

Alphaproteobacteria

Firmicutes

Actinobacteria

Cyanobacteria/ chloroplasts

BetaproteobacteriaGammaproteobacteria

BacteroidetesThermi (Deinococci)Acidobacteria

Deltaproteobacteria

TM-7Chloroflexi

Oxalobacteraceae

Micrococcaceae




NocardiodaceaeSphingomonadaceae (Kaistobacter)Other/unclassified

AcetobacteraceaeSphingomonadaceae

Caulobacteraceae Microcobacteraceae (Agrococcus)

Geodermatophilaceae Methylobacteraceae (Methylobacterium)

Nocardiodaceae (Rhodococcus)Rhodobacteraceae (Rubellimicrobium)

Comamonadaceae Erythrobacteraceae

Chitinophagaceae (Flavolisobacter)Sporichthyaceae

Rhodospirillaceae (Skermanella)Rhizobiales

OxalobacteraceaeChloroplast

BurkholderialesAcetobacteraceae (Roseococcus)

BeijerinckiaceaeNocardiodaceae (Nocardioides)

GeodermatophilaceaeMethylobacteraceae

MicrobacteraceaeGeodermatophilaceae (Modestobacter)

Dietziaceae (Dietzia)Cytophagaceae (Hymenobacter)Cytophagaceae (Adhaeribacter)

ChloroplastBradyrhizobiaceae

Caulobacteraceae (Phenylobacterium)Micrococcaceae

Trueperaceae (Truepera)

Pseudoocardiodaceae (Actinomycetospora)

A.

B.

Top

Inside

*

*

**

*

*

*

*

* Micrococcaceae


Nocardiodaceae


Other/unclassified

Oxalobacteraceae




Micrococcaceae







Alphaproteobacteria

Firmicutes

Actinobacteria




Deltaproteobacteria

TM-7Chloroflexi

Oxalobacteraceae

Micrococcaceae






















A.

B.

Top

Inside

*

*

**

*

*

*

*

*

Outside

A.

B.

14

2009). All of the actinobacterial species that we isolated belong to genera identified in the sequence data set (see the taxa indicated with * in Fig. 4). However, the species successfully cultivated are not the most numerically abundant in the sequence data set. For example, although six of our 13 actinobacterial iso-lates belong to the Corynebacteria, they represent less than 0.3 % of the sequences from the interior sample, and an even smaller fraction of the surface sample. Micrococcales are well-represented in both the sequence data sets and in our collection of isolates. Better understanding how these isolates survive and grow in concrete will therefore give us information about representative bacteria in this unique envi-ronment.

The gene amplification step of this analysis has some biases, so it cannot capture 100% of the total bacte-rial diversity. However, we can use statistical analyses to determine whether our analysis detected the majority of the species present. This rarefaction analysis quantifies the number of taxa observed in sub-sampled data (Fig. 5). If the data has sampled all of the microbial groups present, then the curve of ob-served taxa vs. size of subsample will plateau, indicating that more data does not provide more infor-mation. For both the interior and surface samples, these curves approach a plateau, indicating that we sequenced enough to sample the majority of the taxa in and on concrete. This data was then used to the calculate the Chao1 richness estimator, which uses the number of observed taxa in subsampled data to extrapolate the total number of taxa in a sample (Chao, 1984; Colwell and Coddington, 1994). Based on this estimator, we predict that approximately 1300 taxa are present in both samples. Since we observed

approximately 1100 taxa in these samples, addi-tional sequencing effort might identify additional rare taxa, but would probably not change the overall community composition described here.

The objectives of this phase of the research were to confirm that viable bacteria are present in con-crete, develop a method for extraction of DNA from hardened concrete, and characterize bacte-rial communities in and on concrete. Here, we show that viable bacteria are present in concrete up to a year after pouring and curing, as demon-strated by successful cultivation and isolation of 15 different bacterial strains. We further demon-strate that DNA can be extracted from concrete and is pure enough to be amplified and se-quenced, and that the microbial community in and on these concrete cylinders is dominated by two bacterial phyla, the Actinobacteria and Prote-obacteria. The DNA extraction method can be used to characterize microbial communities in

different types of concrete, in concrete exposed to different environmental conditions, or in corroding concrete.

The DNA sequence analysis also demonstrated that the interior of a concrete structure and its surface are inhabited by different microbes. This suggests that the bacteria in and on concrete are not just remnants

4000 8000 12000 1600000

200

400

600

800

1000

1200

1400

sequences per subsample

obse

rved

OTU

s

TopInside

FIGURE 5. RAREFACTION CURVES. The sequence data was subsampled (ten replicates for each level of sub-sampling) and the average number of operational tax-onomic units (OTUs; analogous to bacterial species) observed in each subsample was plotted. Error bars in-dicate one standard deviation. The lines for both the inside and top sample approach a plateau, indicating that our sequence coverage was sufficient to sample the majority of the taxa present in and on concrete. This figure is reprinted with permission from Maresca, Moser and Schumacher, 2017.

15

of cells that were on the surfaces of the components (or in the water) at the time of mixing, but may in-stead reflect the different kinds of bacteria that survive in slightly different environmental conditions.

3.2 PHASE II: ASR BIOINDICATORS

In phase 2 of this research, we prepared 15 small test cylinders using concrete materials from the Delaware Department of Transpor-tation (DelDOT) known to be highly ASR-reactive, and a parallel series of test cylin-ders that incorporated DelDOT’s standard mitigation against ASR. All cylinders were placed outside on a roof to weather, and one cylinder from each series was removed from the roof every 6-8 weeks and archived (Fig. 6). Three subsamples (5 g each) were collected from each cylinder and DNA was extracted from these. The 16S rRNA genes in these samples were then amplified and sequenced at the University of Delaware Se-quencing and Genotyping Center, and the sequence data was analyzed using Qiime2 software. A total of 30 cylinders (15 from

each series) was analyzed.

3.2.1 Overview of sequencing results

A total of 7.9 million high-quality sequencing reads were obtained from 90 concrete samples (30 cylinders x 3 replicate subsamples of each), as well as the constituent materials and sterile glass beads. We re-moved any taxa observed in the glass beads, which we attribute to laboratory contamination: this is a conservative approach to the analysis, but because the goal was to identify taxa unique to ASR-reactive concrete, it is reasonable. On average, each subsample had 88,000 reads, representing nearly 60,000 in-dividual taxa (Table 4). As in the Phase I study, Proteobacteria are abundant in the bacterial assemblages in concrete; in contrast to the Phase I data, Firmicutes were also abundant, while relatively few Actino-bacteria were observed. When data from all ASR-reactive samples were pooled and compared to the pooled ASR-mitigated samples, it became clear that on the whole, the bacteria in these two kinds of con-crete are quite similar: any gray lines in Fig. 7 indicate bacteria found in both sample types. Because the two concrete batches were made from the same materials, with the only difference being slag in the miti-gated batch, we expected significant similarity, at least in the earlier samples. Some obvious differences between the sample types include some Bacillus and Corynebacterium species that seem to be unique to ASR-reactive samples (green lines in Fig. 7), and Citrobacter and Cloacibacterium species that may be unique to the mitigated samples (blue lines in Fig. 7).

FIGURE 6. EXPERIMENTAL PLAN FOR IDENTIFYING POTENTIAL BIO-INDICATORS FOR ASR IN CONCRETE. Flow chart of sample production, weathering, and analysis.

16

Sam

ple

No.

Rea

ctiv

e/

Miti

gate

d

Sam

ple

date

Avg

erag

e te

mpe

ratu

re1

Prec

ipita

-tio

n2

Replicate 1 Replicate 2 Replicate 3

Reads3 ASVs4 Reads ASVs Reads ASVs

1 ASR reactive 5/2/2013 54°F (24°-90°) 2.69” 75445 59993 84161 68191 82143 66303

2 Mitigated 5/2/2013 54°F (24°-90°) 2.69” 88287 68562 91038 58232 94472 67448

3 ASR reactive 7/15/2013 78°F (56°-95°) 7.37” 81040 63397 88774 62840 81725 64871

4 Mitigated 7/15/2013 78°F (56°-95°) 7.37” 91456 56991 115578 69197 86844 62189

5 ASR reactive 9/4/2013 74°F (53°-90°) 4.46” 77250 56110 95566 60468 88631 62570

6 Mitigated 9/4/2013 74°F (53°-90°) 4.46” 70518 54018 91920 75375 94388 71972

7 ASR reactive 10/23/2013 62°F (39°-87°) 1.92” 81194 65839 100977 69455 86257 55266

8 Mitigated 10/23/2013 62°F (39°-87°) 1.92” 92556 59868 85532 65702 96887 68936

9 ASR reactive 2/22/2014 28°F (3°-66°) 5.43” 73641 26990 76663 59526 75402 57663

10 Mitigated 2/22/2014 28°F (3°-66°) 5.43” 93201 75153 151243 32459 69955 49065

11 ASR reactive 3/31/2014 36°F (8°-69°) 1.73” 75000 61498 85871 60577 92859 56945

12 Mitigated 3/31/2014 36°F (8°-69°) 1.73” 85295 57809 119582 62235 93620 57254

13 ASR reactive 4/20/2014 49°F (21°-83°) 4.04” 90760 70470 86042 68910 68821 33269

14 Mitigated 4/20/2014 49°F (21°-83°) 4.04” 96094 73135 87609 57942 94262 78244

15 ASR reactive 5/20/2014 60°F (35°-84°) 7.38” 79908 64168 97676 78658 85214 71480

16 Mitigated 5/20/2014 60°F (35°-84°) 7.38” 98207 70866 83223 68609 107642 65762

17 ASR reactive 6/19/2014 70°F (49°-93°) 4.99” 55026 43733 93807 74096 94794 75763

18 Mitigated 6/19/2014 70°F (49°-93°) 4.99” 93307 62271 104968 73109 89125 66732

19 ASR reactive 7/15/2014 77°F (55°-93°) 4.74” 86097 62643 87284 61792 66812 51757

20 Mitigated 7/15/2014 77°F (55°-93°) 4.74” 85807 72080 88042 58134 95709 70579

21 ASR reactive 8/17/2014 73°F (56°-91°) 5.50” 92840 59757 91950 63745 96807 62289

22 Mitigated 8/17/2014 73°F (56°-91°) 5.50” 92878 71736 83580 64053 76765 63593

23 ASR reactive 9/17/2014 72°F (47°-92°) 2.95” 91949 56900 82652 53062 89652 63158

24 Mitigated 9/17/2014 72°F (47°-92°) 2.95” 82804 57243 97880 75048 82698 64311

25 ASR reactive 10/19/2014 63°F (39°-83°) 3.46” 84853 38632 95889 74128 81595 64975

26 Mitigated 10/19/2014 63°F (39°-83°) 3.46” 55167 12843 87179 66380 126381 28051

27 ASR reactive 11/19/2014 49°F (19°-73°) 3.32” 78049 54994 70132 50747 83486 35373

28 Mitigated 11/19/2014 49°F (19°-73°) 3.32” 93661 58558 91460 67824 95955 46560

29 ASR reactive 2/17/2015 29°F (2°-53°) 3.92” 84810 42800 86511 60455 119608 55592

30 Mitigated 2/17/2015 29°F (2°-53°) 3.92” 86630 54272 1313 69 84433 55156

17

TABLE 4. SEQUENCE DATA FROM PHASE II EXPERIMENT. Number of reads and sequence variants (analogous to species) identified in the replicates from each sample. Weather conditions (average temperature, temperature range, and total precipitation for the previous month) are also indicated.

1average temperature in degrees Fahrenheit for the month prior to collection. Values in parentheses are the lowest and highest recorded temperature during that period.

2total rainfall in inches during the month prior to collection.

3Number of joined sequences that were analyzed after quality filtering.

4ASVs are “amplicon sequence variants”, groups of sequences that are extremely similar to each other and indicate taxonomic relatedness (analogous to, but more specific than, species). See (Callahan et al., 2017) for additional detail.

3.2.2 Sources of bacteria in concrete

One obvious question, when considering bacteria in or on concrete, is where those bacteria come from. To address this, we extracted DNA from triplicate subsamples of each of the components used to mix the concrete: coarse aggregate, fine aggregate, cement, and slag. The 16S rRNA genes from these samples were amplified and sequenced as for the concrete samples.

The SourceTracker software package (Knights et al., 2011; Hewitt et al., 2013) was used to compare the bacteria identified in the different components of concrete to those identified in the initial samples of the time series (Fig. 8). In the ASR-reactive (“Unmitigated”) samples, most of the bacteria are from an “Un-known” source. Because we did not sample the tap water used to mix the concrete in 2013, we cannot confirm that these bacteria originally came from the water; however, this seems like the most probable source, since the initial t0 samples were never placed outside. In the slag-containing “Mitigated” samples, which should be less ASR-reactive, some of the bacteria seem to come from the slag itself, but the largest fraction of bacteria appears to come from the large aggregate. Although the sand was the component with the highest DNA yield and most bacterial diversity, it appears to contribute very little to the concrete microbiome.

3.2.3 Changes in bacterial assemblages over time

At a very broad level, it is clear that the bacterial assemblages in concrete change over time: the alpha diversity, a metric that incorporates both number of observed taxa and number of individual observations of each taxon, varies as a function of time. The five most abundant bacterial phyla, Proteobacteria, Fir-micutes, Actinobacteria, Bacteroidetes, and Cyanobacteria, are present at fairly consistent levels through-out the time series, with slight upward trajectories (Fig. 9, top row). The next five most abundant phyla appear to vary seasonally, and there are differences between the ASR-reactive (unmitigated) and ASR-mitigated samples (Fig. 9, bottom row).

18

FIGURE 7. COMPARISON OF BACTERIA FOUND IN ASR-REACTIVE AND ASR-MITIGATED CONCRETE. All of the bacte-ria identified in any of the 30 concrete cylinders were quantified. Any taxa present in both types of sample at similar levels are indicated in gray; organisms colored green are more abundant in the ASR-reactive samples, while organisms colored blue are more abundant in the mitigated samples. This figure was made using Metacoder (Foster et al., 2017).

Mitigated

Unmitigated

Bacteria

Proteobacteria

Firmicutes

Cyanobacteria

Bacteroidetes Actinobacteria

Acidobacteria

Alphaproteobacteria

Bacilli

Gammaproteobacteria

Chloroplast

Flavobacteriia

Betaproteobacteria

Cytophagia

Actinobacteria

Sphingobacteriia

Clostridia

Deltaproteobacteria

Bacteroidia[Saprospirae]

Bacillales

Xanthomonadales

Rhodobacterales

Streptophyta

Rhodospirillales

Flavobacteriales

Burkholderiales

Pseudomonadales

Oceanospirillales

Cytophagales

Actinomycetales

Sphingobacteriales

Clostridiales

Enterobacteriales

Rhizobiales

Sphingomonadales

Lactobacillales

Caulobacterales

Alteromonadales

Bacteroidales

Myxococcales

[Saprospirales]

Rickettsiales

Pasteurellales

Xanthomonadaceae

Rhodobacteraceae

Bacillaceae

Comamonadaceae

Pseudomonadaceae

Halomonadaceae

Sphingobacteriaceae

Enterobacteriaceae

BradyrhizobiaceaeEnterococcaceae

Nocardioidaceae

Caulobacteraceae

Moraxellaceae

[Tissierellaceae]

Sphingomonadaceae

Chitinophagaceae

Staphylococcaceae

Micrococcaceae

Oxalobacteraceae

Corynebacteriaceae

[Weeksellaceae]

mitochondria

Methylobacteriaceae

Streptococcaceae

Sinobacteraceae

Pasteurellaceae

Stenotrophomonas

Bacillus

Pseudomonas

Halomonas

Acinetobacter

Erwinia

Sphingomonas

Sediminibacterium

Staphylococcus

Corynebacterium

Enterococcus

Cloacibacterium

Citrobacter

19

3.2.4 Differences between ASR-reactive and ASR-mitigated con-crete

An ideal bio-indicator bacterium would be present in only ASR-reactive or ASR-mitigated samples at the end of the time series, would clearly change over time, and would be abundant enough to be reliably detectable. No species matching all these criteria have yet been identi-fied. However, there are clear differ-ences in the composition of and sea-sonal variability within the microbial assemblages in the two sample types. We first examined the species identified in Fig. 7 as being different in the ASR-reactive or unreactive samples, and plotted their abundance over time in all samples (Fig. 10). In most cases, although there are differences between ASR-reactive and mitigated samples that explain the results observed in Fig. 7, the differences at the end of the time series are not large enough to make these species reliable bio-indicators.

Acidobacteria [Thermi] Fusobacteria Planctomycetes Chloroflexi

Proteobacteria Firmicutes Actinobacteria Bacteroidetes Cyanobacteria

Jul 1

3

Jan

14

Jul 1

4

Jan

15

Jul 1

3

Jan

14

Jul 1

4

Jan

15

Jul 1

3

Jan

14

Jul 1

4

Jan

15

Jul 1

3

Jan

14

Jul 1

4

Jan

15

Jul 1

3

Jan

14

Jul 1

4

Jan

150.001

0.100

0.001

0.100

Months

Rel

ativ

e ab

unda

nce

Mitigation status

mitigated

unmitigated

FIGURE 9. CHANGES IN PHYLUM ABUNDANCE OVER TIME. Relative abundance of each phylum is plotted as a func-tion of date (month and year of collection are indicated in the x-axis labels). Some phyla, such as Cyanobacteria, appear to vary seasonally, while others such as the Bacteroidetes stay constant over time. Each point represents a species belonging to the indicated phylum (e.g. Proteobacteria); lines indicate averages of the phylum as a whole.

Bacteria source Concrete composition

Mitigated

Unm

itigated

0%

25%

50%

75%

100%

0%

25%

50%

75%

100%

Months

Per

cent

Source

FlyAsh

Gravel

Powder

Sand

Unknown

Water

Source of Bacteria in Concrete Cylinders

FIGURE 8. SOURCES OF BACTERIA IN CONCRETE. The bacteria identi-fied in the concrete were compared to bacteria identified in each of the components, and the proportion of bacteria originating in each component, or from “unknown” sources, was predicted using Source-Tracker.

20

However, other species did have different patterns of variability over time. For example, the Aerocococ-cus species appear to be similar for the first few months of the time series, as expected, since the two se-ries were made with identical materials (the ASR-mitigated batch replaced half of the cement with slag, but that was the only change). In ASR-mitigated concrete, Aerococcus spp. appear to increase in January, peak in April, and decrease as the weather warms (Fig. 11). In ASR-reactive concrete, the same species decrease in January, but begin to increase in March and slowly increase through October. The Lactobacil-lus and Exiguobacterium spp., similarly, appear to be present at different abundances and respond differ-ently to weather (Fig. 10). The differences between ASR-reactive and ASR-mitigated samples at the end of the time series make species within these two genera attractive as possible bio-indicators of ASR.

FIGURE 10. CHANGES OVER TIME IN THE DIFFERENTIALLY ABUNDANT GENERA IDENTIFIED IN FIGURE 6. Abundance of gen-era (labeled “(G)” above) or spe-cies (S) is plotted as a function of collection date, and all repli-cates are plotted individually. Each point represents an obser-vation of a species within the in-dicated group (e.g. Bradyrhizo-bium), and lines indicate aver-ages of the group as a whole.

21

More useful than either of these are the Staphylococcus spp. (Fig. 11). These are consistently observed throughout the time series, suggesting that they are abundant enough to be reliably detectable. The sea-sonal dynamics in the ASR-reactive and ASR-mitigated samples are different, but the Staphylococci appear to increase over time in the mitigated samples, while they stay at approximately the same level in the un-mitigated samples. Other genera currently under consideration as bio-indicators are Aeromicrobium, Chi-tinophaga, Sediminibacterium, and Xenococcus, all of which have different abundances in the two sample types at the final time point.

FIGURE 11. CHANGES OVER TIME IN SELECTED DIFFERENTIALLY ABUNDANT GENERA. Relative abundance of each genus is plotted as a function of date (month and year of collection are indicated in the x-axis labels). Each point represents an observation of a species within the indicated group (e.g. Bradyrhizobium), and lines indicate aver-ages of the group as a whole. Genera were selected based on apparent differences at the final time point.

4 CONCLUSIONS

4.1 GENERAL CONCLUSIONS

Based on the results from our Phase I (proof-of-principle) and phase II (time series) experiments, we have demonstrated that viable bacteria are present in concrete, and that DNA can be extracted from concrete and used to interrogate the concrete “microbiome.” We have further shown that most of the bacteria originate from the solid materials used in the concrete mix, and that the bacterial assemblages in con-crete change over time. In a 2-year time-series comparing the bacteria in ASR-reactive and mitigated

22

concrete, the differences between the two sample types are minimal in the early time points, but in-crease over time. We observe seasonable variations in the bacterial communities in both sample types, and those variations are not identical. These results suggest that the bacterial communities in concrete depend on the composition of the concrete and are dynamic, changing in response to environmental con-ditions. Consequently, development of bacterial bio-indicators for specific types of chemical damage in concrete is feasible.

4.2 CAVEATS

Although we are hopeful about developing bio-indicators for early detection of ASR-induced damage, this research is still in an early stage. Some technical concerns include the possible presence of “relic DNA”, variability between samples, regional variations in materials used, heterogeneity within samples, and the extremely low biomass present in concrete. These will need to be resolved before bio-indicators can be used as a practical method for early detection of ASR or any other chemical degradation of concrete.

Concrete production is a very local industry: both coarse and fine aggregates, as well as the water used in the mix, are obtained from local sources. Since the majority of the bacteria identified in our samples came from the components used in the concrete mix (Fig. 8), it is possible that the bacterial communities in concrete produced in different regions, with different starting materials, could be quite different. Since concrete heats up as it cures, and since cured concrete is alkaline and dry, it is possible that these similari-ties result in similar bacterial communities, even if the bacteria in the components might be quite differ-ent. To determine whether there are regional differences in bacterial communities in concrete, or whether the similarity in concrete chemical and material properties results in a general “concrete micro-biome,” bacterial communities from concrete from different regions would need to be characterized. If there are significant regional differences in concrete “microbiomes”, then unique bio-indicators would need to be identified in every region.

Potentially of even greater concern than variation between regional samples is the heterogeneity we ob-serve within samples. In some cases, subsamples of the same concrete cylinder are as different from each other as they are from other cylinders. Because concrete is solid, and water flow through the matrix only occurs after the concrete has cracked, there is no mixing within the sample, and the intra-sample variabil-ity is very high. This has also been a consistent problem with analysis of microbial communities in soil samples (Lombard et al., 2011). As a result of this great heterogeneity, distinguishing between the species that change with respect to age or environmental variables and those that change with respect to loca-tion within a sample can be difficult. In soil systems, this problem can be solved by large enough sample sizes and by combining replicate samples, rather than analyzing them separately (Lombard et al., 2011). We anticipate that a similar approach will work here. The 5-g samples we collected have enough DNA to analyze, but it is very close to the limit of detection. Larger samples (10g or greater) might yield improved results, and data from replicate samples can be pooled to yield a more representative picture of the mi-crobes found in each sample.

Another concern in this system is distinguishing between DNA from live cells, which can respond to changing environmental conditions, and DNA from dead cells. DNA is fairly stable even outside of the cell, and “relic DNA” can cause problems in this type of DNA-dependent analysis (Carini et al., 2016). In some cases, extracellular DNA can be degraded prior to extraction of DNA from cells, to prevent analysis of the relic DNA. However, because the crushing step in our protocol may break some cells open and release

23

DNA, pre-treatment of extracellular DNA seems im-practical. Instead, we are developing a method that uses confocal microscopy to visualize bacterial cells in concrete. We have successfully visualized E. coli on the surface of a small concrete chip (Fig. 12). If visu-alization of bacteria in and on concrete can be done using fluorescence microscopy, live/dead assays or activity assays that utilize dyes whose fluorescence properties change in response to biological activity (such as changes in pH or redox potential) can be used on these samples. This would allow us to deter-mine how many of the cells in concrete are intact and active relative to those that have damaged membranes or are no longer active.

4.3 PROPOSED BIOINDICATORS AND FUTURE WORK

Our goal was to identify the bacteria present in ASR-reactive or mitigated concrete as the ASR damage developed and concrete weathered, then identify spe-cific types of bacteria present at early stages of ASR damage. An ideal bio-indicator bacterium would be present in only ASR-reactive or ASR-mitigated samples at the end of the time series, would clearly change over time, and would be abundant enough in all samples to be reliably detectable. In this data set, we have identified five candidate bio-indicators: Staphylococcus, Aeromicrobium, Chitinophaga, Sediminibac-terium, and Xenococcus.

Additional testing will be required to determine the suitability of these organisms as bio-indicators in the field. These tests will include quantitative PCR (qPCR) against the 16S rRNA genes from these species in the existing DNA samples, to (a) confirm the results from the amplicon sequencing, and (b) more quanti-tatively determine the abundance of these species, since amplicon sequencing can be used for relative abundance but is not a reliable indicator of absolute abundance. The same qPCR procedure can then be used to quantify abundance of these species in ASR-damaged and undamaged samples obtained from local departments of transportation. Some samples have already been obtained from West Virginia and Delaware DOTs.

If the concerns outlined above can be overcome, it may also be possible to identify bioindicators for other types of damage. Additional work should also be done to determine whether different bioindicators would be necessary for reinforced concrete or for concrete with additives such as polymers that could potentially provide organic carbon to microbes.

FIGURE 12. E. COLI ON CONCRETE SURFACE. Con-crete was incubated in a solution with E. coli over-night, then stained with Syto59, a DNA stain. Both reflectance data (gray) and fluorescence (red) was collected in a Z-stack and used to reconstruct a 3-di-mensional image of the cells on the concrete sur-face.

24

5 REFERENCES

Amini, K. and Kraatz, H.-B. (2014) Recent advances and developments in monitoring biological agents in water samples. Rev. Environ. Sci. Bio/Technology 14: 23–48.

Arndt, R.W., Schumacher, T., Algernon, D., and Kee, S.-H. (2011) Strategies for maintenance of highway bridges in the US-with the support of nondestructive testing and structural health monitoring. Bautechnik 88: 793–804.

Aronesty, E. (2013) Comparison of sequencing utility programs. Open Bioinforma. J. 7: 1–8.

Callahan, B.J., McMurdie, P.J., and Holmes, S.P. (2017) Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. ISME J. 11: 2639–2643.

Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-Lyons, D., Huntley, J., Fierer, N., et al. (2012) Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J. 6: 1621–4.

Carini, P., Marsden, P.J., Leff, J.W., Morgan, E.E., Strickland, M.S., and Fierer, N. (2016) Relic DNA is abundant in soil and obscures estimates of soil microbial diversity. Nat. Microbiol. 2: 16242.

Cattaneo, C., Smillie, D.M., Gelsthorpe, K., Piccinini, A., Gelsthorpe, A.R., and Sokol, R.J. (1995) A simple method for extracting DNA from old skeletal material. Forensic Sci. Int. 74: 167–174.

Chao, A. (1984) Nonparametric Estimation of the Number of Classes in a Population. Scand. J. Stat. 11: 265–270.

Chen, W.-F. and Liew, J.Y.R. eds. (2003) The civil engineering handbook 2nd ed. CRC Press, New York.

Colwell, R.K. and Coddington, J.A. (1994) Estimating terrestrial biodiversity through extrapolation. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 345: 101–18.

Delatte, N.J. (2001) Lessons from Roman Cement and Concrete. J. Prof. Issues Eng. Educ. Pract. 127: 109–115.

Demkina, E. V., Soina, V.S., Registan, G.I.E., and Zvyagintsev, D.G. (2000) Reproductive resting forms ofArthrobacter globiformis. Microbiology 69: 309–313.

Duckworth, A., Grant, S., Grant, W., Jones, B., and Meijer, D. (1998) Dietzia natronolimnaios sp. nov., a new member of the genus Dietzia isolated from an east African soda lake. Extremophiles 2: 359–66.

Epstein, S. (2013) The phenomenon of microbial uncultivability. Curr. Opin. Microbiol. 16: 636–642.

Foster, Z.S.L., Sharpton, T.J., and Grünwald, N.J. (2017) Metacoder: An R package for visualization and manipulation of community taxonomic diversity data. PLOS Comput. Biol. 13: e1005404.

Giannantonio, D.J., Kurth, J.C., Kurtis, K.E., and Sobecky, P.A. (2009) Molecular characterizations of microbial communities fouling painted and unpainted concrete structures. Int. Biodeterior. Biodegradation 63: 30–40.

Gilbert, J.A., Jansson, J.K., and Knight, R. (2014) The Earth Microbiome project: successes and aspirations.

25

BMC Biol. 12: 69.

Harwood, V.J., Staley, C., Badgley, B.D., Borges, K., and Korajkic, A. (2014) Microbial source tracking markers for detection of fecal contamination in environmental waters: relationships between pathogens and human health outcomes. FEMS Microbiol. Rev. 38: 1–40.

Hewitt, K.M., Mannino, F.L., Gonzalez, A., Chase, J.H., Caporaso, J.G., Knight, R., and Kelley, S.T. (2013) Bacterial Diversity in Two Neonatal Intensive Care Units (NICUs). PLoS One 8: e54703.

Hugenholtz, P., Pitulle, C., Hershberger, K.L., and Pace, N.R. (1998) Novel division level bacterial diversity in a Yellowstone hot spring. J. Bacteriol. 180: 366–76.

Kim, S.B., Nedashkovskaya, O.I., Mikhailov, V. V, Han, S.K., Kim, K.-O., Rhee, M.-S., and Bae, K.S. (2004) Kocuria marina sp. nov., a novel actinobacterium isolated from marine sediment. Int. J. Syst. Evol. Microbiol. 54: 1617–20.

Knights, D., Kuczynski, J., Charlson, E.S., Zaneveld, J., Mozer, M.C., Collman, R.G., et al. (2011) Bayesian community-wide culture-independent microbial source tracking. Nat. Methods 8: 761–763.

Kouduka, M., Suko, T., Morono, Y., Inagaki, F., Ito, K., and Suzuki, Y. (2012) A new DNA extraction method by controlled alkaline treatments from consolidated subsurface sediments. FEMS Microbiol. Lett. 326: 47–54.

Kuczynski, J., Stombaugh, J., Walters, W.A., González, A., Caporaso, J.G., and Knight, R. (2012) Using QIIME to analyze 16S rRNA gene sequences from microbial communities. Curr. Protoc. Microbiol. Chapter 1: Unit 1E.5.

Lombard, N., Prestat, E., van Elsas, J.D., and Simonet, P. (2011) Soil-specific limitations for access and analysis of soil microbial communities by metagenomics. FEMS Microbiol. Ecol. 78: 31–49.

Loreille, O.M., Diegoli, T.M., Irwin, J.A., Coble, M.D., and Parsons, T.J. (2007) High efficiency DNA extraction from bone by total demineralization. Forensic Sci. Int. Genet. 1: 191–5.

Maresca, J.A., Moser, P., and Schumacher, T. (2017) Analysis of bacterial communities in and on concrete. Mater. Struct. 50: 25.

Marine, S.C., Pagadala, S., Wang, F., Pahl, D.M., Melendez, M. V, Kline, W.L., et al. (2015) The growing season, but not the farming system, is a food safety risk determinant for leafy greens in the mid-Atlantic region of the United States. Appl. Environ. Microbiol. 81: 2395–407.

Sarafin, Y., Donio, M.B.S., Velmurugan, S., Michaelbabu, M., and Citarasu, T. (2014) Kocuria marina BS-15 a biosurfactant producing halophilic bacteria isolated from solar salt works in India. Saudi J. Biol. Sci. 21: 511–9.

Shimkets, L.J. (2013) Prokaryotic Life Cycles. In, Rosenberg,E., DeLong,E.F., Lory,S., Stackebrandt,E., and Thompson,F. (eds), The Prokaryotes: Prokaryotic Communities and Ecophysiology. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 317–336.

Thompson, L.R., Sanders, J.G., McDonald, D., Amir, A., Ladau, J., Locey, K.J., et al. (2017) A communal catalogue reveals Earth’s multiscale microbial diversity. Nature 551: 457.

Touma, W.E., Fowler, D.F., and Carrasquillo, R.L. (2001) Alkali-silica reaction in portland cement concrete:

26

testing methods and mitigation alternatives Austin, TX.

Wegen, G. Van der, Polder, R.B., and Breugel, K. Van (2012) Guideline for service life design of structural concrete: A performance based approach with regard to chloride induced corrosion. Heron 57: 3.

Wei, S., Sanchez, M., Trejo, D., and Gillis, C. (2010) Microbial mediated deterioration of reinforced concrete structures. Int. Biodeterior. Biodegrad. 64: 748–754.

Wolin, E.A., Wolin, M.J., and Wolfe, R.S. (1963) Formation of Methane by Bacterial Extracts. J. Biol. Chem. 238: 2882–2886.

Yoon, J.-H., Lee, C.-H., and Oh, T.-K. (2006) Nocardioides lentus sp. nov., isolated from an alkaline soil. Int. J. Syst. Evol. Microbiol. 56: 271–5.

Young, M., Artsatbanov, V., Beller, H.R., Chandra, G., Chater, K.F., Dover, L.G., et al. (2010) Genome sequence of the Fleming strain of Micrococcus luteus, a simple free-living actinobacterium. J. Bacteriol. 192: 841–60.

Zhou, J., Bruns, M.A., and Tiedje, J.M. (1996) DNA recovery from soils of diverse composition. Appl. Environ. Microbiol. 62: 316–22.

27

6 APPENDICES

APPENDIX 1: INVITED SEMINARS PRESENTING THE FUNDED WORK.

1. J.A. Maresca. Hard Microbiology: The concrete microbiome. University of Texas at San Antonio, South Texas Center for Emerging Infectious Diseases. February 2, 2018.

2. J.A. Maresca. Hard Microbiology: Bacteria in concrete. University of Maryland, Department of Civil and Environmental Engineering. December 8, 2017.

3. J.A. Maresca. Bioinformatics in civil engineering: Identifying bacterial bio-indicators for damaged con-crete. University of North Carolina, Charlotte, Department of Bioinformatics. October 20, 2017.

4. J.A. Maresca. Civil microbiology: Bacteria in natural and engineered systems. University of Delaware, Department of Civil and Environmental Engineering. May 13, 2016.

5. J.A. Maresca. Hard Microbiology: Bacteria in concrete. Towson University, Department of Physics, As-tronomy, and Geosciences, Towson, MD. April 8, 2016.

6. J.A. Maresca. Hard Microbiology: Bacteria in concrete. Michigan State University, Department of Geo-logical Sciences, East Lansing, MI. November 20, 2015.

7. J.A. Maresca. Hard Microbiology: Bacteria in concrete. Wentworth Institute of Technology, Department of Civil Engineering and Technology. November 4, 2015.

8. J.A. Maresca. Hard Microbiology: Bacteria in concrete. Villanova University, Department of Civil and Environmental Engineering. October 23, 2015.

28

APPENDIX 2: CONFERENCE PRESENTATIONS OF THE FUNDED WORK

1. J.A. Maresca*. Urban geomicrobiology: Microbial communities in and on concrete. Geological Society of America Annual Meeting, September 2016. (Invited talk.)

2. J.A. Maresca*, J.L. Keffer. Microbial communities in concrete and their potential application in ASR-damaged concrete. Mid-Atlantic Transportation Sustainability Center – Region 3 University Transpor-tation Center Annual Meeting, August 2016. (Poster presentation and invited talk.)

3. A.K. Treglia*, J.L. Keffer, J.A. Maresca. Using Microbial Populations in Concrete as Bio-Indicators of Alkali-Silica Reaction. American Society of Microbiology General Meeting/Microbe, June 2016. (Poster presentation)

4. J.A. Maresca*, K. Zhang, T. Schumacher. Naturally occurring bacterial communities in and on con-crete. American Concrete Institute, Research In Progress session. November 2015. (Oral presenta-tion.)

5. K. Zhang, J.A. Maresca. Microbial communities in concrete and their potential application in ASR-damaged concrete. Mid-Atlantic Transportation Sustainability Center – Region 3 University Transpor-tation Center Annual Meeting, August 2015. (Poster presentation.)

6. P. Moser*, K. Zhang, T. Schumacher, J.A. Maresca. Microbially mediated detection of damaged con-crete. International Conference on Ecology and Transportation, September 2015. (Poster presentation)

7. J.A. Maresca*, K. Zhang. Hard microbiology: The bacteria in concrete. American Phytopathological Soci-ety, Potomac Chapter, March 2015. (Invited talk.)

8. J.A. Maresca*, K. Zhang, P. Moser, J. Moore, T. Schumacher. Bacterial indicators of ASR-induced dam-age. Corvallis Workshop on Cement-Based Materials, July 2014. (Invited talk)

9. J.A. Maresca*, P. Moser, M.K. Sutter, F. Rajabipour, T. Schumacher. Bio-indicators of ASR-induced dam-age. American Concrete Institute, Minneapolis MN, April 2013. (Oral presentation.)

29

APPENDIX 3. PUBLICATION: MARESCA ET AL., MATERIALS AND STRUCTURES 2017

ORIGINAL ARTICLE

Analysis of bacterial communities in and on concrete

Julia A. Maresca . Paul Moser .

Thomas Schumacher

Received: 23 February 2016 / Accepted: 6 July 2016 / Published online: 10 August 2016

� RILEM 2016

Abstract Bacteria are known to catalyze degrada-

tion of concrete, and have more recently been used to

repair micro-cracks in or form protective biofilms on

cement mortar. However, the microbial communities

in and on concrete under ordinary weathering condi-

tions have not been characterized, in part because of

difficulty in extracting DNA from inside concrete

specimens. Here, we report a method for extraction of

nucleic acids directly from hardened concrete. Using

this method and classical cultivation methods, we

demonstrate that most bacteria in or on concrete

belong to two taxonomic groups, that the bacterial

diversity is similar on the concrete surface and in the

interior, and that many bacteria in and on concrete are

related to microbes found in other dry, saline, or

alkaline environments. This method lays the founda-

tion for the creation of bioindicators for concrete and

may open new avenues for the fields of non-destruc-

tive evaluation and assessment of concrete structures.

Keywords Bacteria � Concrete � Bioindicator � Non-destructive evaluation

1 Introduction and motivation

1.1 Motivation

Worldwide, the concrete civil infrastructure network

is aging and deteriorating, and agencies need inex-

pensive, effective and objective inspection techniques

to make informed decisions with respect to mainte-

nance and repair. In particular, early detection of

damage is crucial, as deterioration processes tend to

accelerate with time, along with the costs for rehabil-

itating a structure. Currently, condition assessments

are mostly performed by using visual inspection

methods, which can be subjective [1]. Non-destructive

evaluation techniques are becoming more popular, but

can be expensive due to the specialized equipment and

expertise needed. Structural health monitoring allows

for long-term collection of data using sensor networks

attached to a structure. The challenge with this

approach lies in the extensive analysis required to

interpret the large amounts of continuously-produced

data. Hence, there is a need for an inspection technique

that is non-invasive, inexpensive, easy to deploy, and

can detect conditions that may lead to deterioration

early on.

Electronic supplementary material The online version ofthis article (doi:10.1617/s11527-016-0929-y) contains supple-mentary material, which is available to authorized users.

J. A. Maresca (&) � P. Moser

Department of Civil and Environmental Engineering,

University of Delaware, Newark, DE 19716, USA

e-mail: [email protected]

T. Schumacher

Department of Civil and Environmental Engineering,

Portland State University, Portland, OR 97201, USA

Materials and Structures (2017) 50:25

DOI 10.1617/s11527-016-0929-y

http://orcid.org/0000-0002-3955-1585

http://dx.doi.org/10.1617/s11527-016-0929-y

http://crossmark.crossref.org/dialog/?doi=10.1617/s11527-016-0929-y&domain=pdf

http://crossmark.crossref.org/dialog/?doi=10.1617/s11527-016-0929-y&domain=pdf

Bioindicators are commonly used in water quality

analysis [2] and food safety assessments [3] to identify

contamination and/or assess risk, and we believe there

is great potential for their application to non-destruc-

tive evaluation of concrete. The presence or activity of

particular species of bacteria may indicate conditions

that lead to deterioration mechanisms such as alkali–

silica reaction (ASR), carbonation, or corrosion of the

embedded reinforcing steel. To assess the potential for

use of bioindicators in non-destructive evaluation, we

must first confirm that viable bacteria are present in

concrete, then develop a method for identifying these

bacteria. The first steps toward identification of

bioindicators for concrete are thus cultivation of

representative bacteria from concrete, the develop-

ment of a method for DNA extraction from concrete,

and subsequent identification of microbial communi-

ties that naturally occur in and on concrete.

1.2 Background

In recent years, the effects of microbes on concrete

degradation [4–6] and repair [7–9] have been charac-

terized extensively. Microbial communities that

degrade concrete have been characterized in detail in

concrete sewers, where sulfur metabolism contributes

to the acid-catalyzed dissolution of the structure

[10, 11], and on submerged structures, where similar

mechanisms operate [12, 13]. Next-generation

sequence analysis has recently been applied to char-

acterize surficial biofilms on these structures, and a

variety of sulfur-metabolizing, acid-tolerant microbes

were identified in these biofilms [14–17].

Additionally, bacterial precipitation of calcium

carbonate (CaCO3) has been tested as a mechanism

for repairing small cracks in concrete, using urea

((NH2)2CO) or calcium lactate as substrates [8, 18],

and encasing the bacteria or bacterial spores in silica

gel or clay [8, 19]. Other work using microbial

biofilms to prevent transport of ions into and out of

cement mortar has shown that these surface-associated

cells can increase concrete service-life [20–22]. Thus,

bacteria are already known to play roles in both

degradation and protection of concrete.

However, little to nothing is known about the

bacteria that live in and on ordinary concrete: the

population dynamics, the heterogeneity within and

between different kinds of concrete, how the microbes

survive the stresses imposed by the salty, alkaline

environment, or even what kinds of microbes they are.

In order to examine the composition of the commu-

nities of bacteria that develop in and on concrete, a

cultivation-independent method for community anal-

ysis is necessary. To further characterize how the

members of this community have adapted to the

extreme conditions present in concrete, it is necessary

to cultivate members of this community in the

laboratory.

Here, we describe the isolation of 15 bacterial

strains for laboratory investigation of the physiology

of survival in concrete. We additionally demonstrate

efficient extraction of DNA directly from concrete,

and use high-throughput sequencing and analysis of

the 16S rRNA gene to describe the composition of the

microbial community in concrete exposed outside

for *12 months. In the future, this cultivation-inde-

pendent analytical method can potentially be used to

monitor changes over time, or even as a proxy for

diagnosis of potential degradation risks.

2 Objectives

The objectives of this study were to confirm the

presence of viable bacteria in concrete, develop a

method for the extraction of DNA from hardened

concrete, and characterize bacterial communities in

and on concrete.

3 Materials and methods

3.1 Preparation and weathering of concrete

To assess whether ordinary pavement concrete could

be colonized by microbes, concrete cylinders

(152 9 305 mm) were cast in December 2011 using

a Delaware Department of Transportation (DelDOT)

Class B mix design (Table 1), which is typically used

for concrete pavements. Crushed coarse aggregates

had a maximum size of 19 mm.

After casting, the cylinders were cured in their

plastic forms for 2 weeks, then uncapped and kept in a

storage room at room temperature. In February 2012,

two cylinders were placed on top of a University

building in Newark, Delaware (USA) (Fig. 1a),

exposed to weather and minimal human interaction

(Fig. 1b). After 12 months of exposure, samples for

25 Page 2 of 13 Materials and Structures (2017) 50:25

cultivation and DNA analysis were collected. Prior to

sampling, cylinders were moved from the rooftop to a

pre-sterilized sampling area. All tools were sterilized

with 95 % ethanol. The top surface of one cylinder

was struck using a 3-in. masonry chisel and hammer

until surface chips or larger chunks spalled off

(Fig. 1c, d).

3.2 Scanning electron microscopy with energy

dispersive X-ray spectroscopy (SEM–EDS)

SEM–EDS was used to quantify the elemental com-

position of the concrete. The concrete cylinder was

sliced into *8 mm-thick horizontal sections, then

sections from the middle of the cylinder were further

cut into pieces approximately 1 9 1 cm. The samples

were attached to aluminum stubs and coated with

Table 1 Specified requirements for DelDOT Class B concrete

mix

Characteristic Specification

Design compressive strength, fc’ 20 MPa (2900 psi)

Design cement content (minimum) 334 kg/m3 (560 lb/yd3)

Design water-to-cement ratio, w/c 0.45

Required air content 4–7 %

Required slump 50–100 mm (2–4 in.)

Required admixtures (ASTM 2013) A, D, E, F, G

Fig. 1 Exposure of cylinders to weather and subsequent

sampling for cultivation, DNA analysis, and SEM–EDS.

a. Concrete cylinders were placed on the roof of a university

building. b. Cylinders were exposed to normal weather

conditions, including wind and precipitation, for 1 year.

c. The cylinder was broken for sampling from the inside for

both cultivation and DNA analysis. d. Slices were also taken forSEM–EDS analysis

Materials and Structures (2017) 50:25 Page 3 of 13 25

3 nm of carbon in a Leica EMACE600 and kept under

vacuum overnight prior to SEM analysis. Three

independent samples were analyzed, and for each,

15–20 2 9 2 mm fields of view were imaged and

analyzed. Elemental analysis was performed with an

Oxford Instruments INCAx-act detector attached to a

Hitachi S-4700 field-emission scanning electron

microscope (FE-SEM).

3.3 Bacterial growth conditions and isolation

and identification of bacterial cultures

Individual chips from the top and side surfaces or

internal areas of the concrete cylinders were collected

in sterile 50 mL centrifuge tubes or WhirlPak bags.

For cultivation, TE (10 mM Tris, 1 mM EDTA) was

added to the concrete pieces in centrifuge tubes. The

samples were vortexed for 10 s and 25 lL of the

buffer was spread on solid CM-A medium, a minimal,

slightly alkaline medium developed for this study.

CM-A contains (per liter), 1 g NaHCO3, 1.44 g

ammonium acetate, 0.12 g MgSO4�7H2O, 0.14 g

KH2PO4, 0.17 g K2HPO4, 1 mL eachWolfe’s vitamin

and trace minerals solutions [23], and 11 mL phos-

phate buffer (62.7 g KH2PO4 L-1, 1.27 g K2HPO4 -

L-1, pH 8.5). The pH of the medium was adjusted to

8.0. Additional samples were gently sonicated to

remove cells from concrete pieces, and buffer was

transferred to the media in the same fashion. The

plates were incubated at room temperature for

2 weeks. Individual colonies were restreaked onto

fresh CM-A media until axenic. Isolation was verified

by microscopy and 16S rRNA sequencing. Once

isolated, strains were archived at -80 �C in 10 %

glycerol.

Colonies were scraped from solid media into TE

and collected by centrifugation. DNA was extracted

by bead-beating in extraction buffer (200 mM NaCl,

200 mM Tris, 20 mM EDTA, pH 8.0), followed by

chloroform extraction and ethanol precipitation. DNA

was resuspended in sddH2O and stored at-20 �C until

use.

The 16S rRNA genes of the isolates were PCR-

amplified using primers 8F and 1492R [24] and

sequenced at the University of Delaware Sequencing

and Genotyping Facility. Sequences were compared to

the non-redundant nucleotide database at NCBI using

BLASTN [25] and were assigned to a species based on

percent similarity.

3.4 Extraction of genomic DNA from concrete

Samples were collected from two sections (top, inside)

of each concrete cylinder. After crushing with mortar

and pestle (sterilized by autoclaving), 5 g pulverized

concrete was transferred into individual 50 mL cen-

trifuge tubes.

3.4.1 DNA extraction methods attempted

DNA extraction from concrete was initially attempted

using the MoBio PowerSoil and PowerMax DNA

extraction kits (MoBio, Carlsbad, CA, catalog nos.

12888 and 12988, respectively), but no genomic DNA

was obtained.Amethod used for extraction ofDNA from

endolithic communities was also attempted [26]. Briefly,

pulverized concrete (1–2 g) was suspended in 5 mL

buffer composed of 200 mM NaCl, 200 mM Tris, and

20 mMEDTA (pH * 8.0). Zirconium beads (0.1 mM),

2 mL 20 % SDS, and 5 mL chloroform were added and

cells were vortexed at maximum speed for 3 min. Lysate

was centrifuged for *5 min and the aqueous phase was

transferred to a fresh centrifuge tube, re-extractedwith an

equal volume of chloroform, then precipitated with 1

volume cold isopropanol and 0.1 volume 3 M ammo-

nium acetate. The DNA pellet was then washed with

70 % ethanol, air-dried, and resuspended in TE (10 mM

Tris, 1 mM EDTA, pH 8.0). No genomic DNA was

obtained using this method.

3.4.2 DNA extraction protocol

Each 5-g sample was washed with 20 mL of 0.5 M

EDTA, followed by vortexing and centrifugation, to

remove divalent cations from the concrete powder [27].

Thepowderwas then suspended in15 mLof lysis buffer

(40 mM EDTA, 50 mM Tris pH 8.3, 0.73 M sucrose)

with 15 mg of lysozyme per gram of concrete powder

and incubated at 37 �Cwith gentle agitation for 30 min.

Next, 100 lL of proteinase K (20 mg mL-1) and

1.5 mL of 20 % sodium dodecyl sulfate (SDS) were

added and the solution was incubated at 56 �C for

2.5–3 h with gentle agitation. Membranes, proteins and

cell debris were removed by extraction with 20 mL

chloroform. To remove additional non-nucleic acid

material, 1 mL of 1.95 M sodium acetate was added to

the aqueous phase of each sample after extraction and

mixed vigorously for 30 s [27]. The sample was then

centrifuged and the aqueous phase was transferred to a


fresh centrifuge tube and re-extracted with 0.8 volume

of chloroform. DNA was precipitated with 1 volume of

isopropanol and 0.1 volume 3 M sodium acetate. The

DNA pellet was washed with 70 % ethanol, and

genomic DNA was then re-suspended in 25 lL of

autoclaved ddH2O and stored at-20 �C.To confirm that biomass is not lost during the

EDTAwash, DNAwas extracted from the EDTA used

to wash the concrete using the MoBio PowerWater

DNA extraction kit (MoBio catalog no. 14900).

3.5 Sequence analysis

Genomic DNA was sent to Research and Testing

Laboratories (Lubbock, TX) for 16S rRNA gene

amplification and sequencing. Primers for amplifica-

tion of the V3–V5 region were 515F and 806R [28]

with Illumina-specific primers and barcodes. Sequenc-

ing was performed on an Illumina MiSeq; both

amplification and sequencing were done by Research

and Testing Laboratories (Lubbock, TX).

Paired raw reads were joined, quality-filtered and

assigned to taxa using Qiime pipelines [29, 30].

Operational taxonomic units (OTUs) were defined as

clusters sharing 97 % sequence identity. Alpha and

beta diversity parameters were calculated using stan-

dard Qiime pipelines [31].

4 Results and discussion

4.1 Concrete characteristics

An analysis using SEM/EDS was performed on small

specimens of the weathered concrete to quantify the

elemental composition (Fig. 2). Oxygen was the most

abundant element, followed by Si, Al, Ca, K and Na.

Small amounts (\1 % each) of Mg, Fe, and S were

also detected in some areas of these samples. Because

the samples were carbon-coated, carbon content could

not be analyzed in this system. The Fe and S could

potentially serve as electron donors or acceptors for

microbial energy metabolism, and have been shown to

do so in corroding concrete sewer pipes [10, 32].

4.2 Bacterial strains isolated

Fifteen bacterial strains were isolated from the surface

and inside of the concrete samples. Greater diversity

was initially observed on solid media spread with

samples from surfaces than interior samples (Table 2).

Individual colonies were transferred to fresh solid

medium until the culture contained only one kind of

organism, as assessed by microscopy and 16S rRNA

amplification and sequencing.

Nearly all of the bacteria isolated from the concrete

test samples are Actinobacteria, primarily Arthrobacter

spp. (Table 3). One representative each of Alphapro-

teobacteria (aParacoccus sp.) andGammaproteobacteria

(Pseudoalteromonas) were also isolated. Most of these

strains are brightly colored due to the synthesis of

carotenoid pigments (data not shown), and grow readily

on both solid and liquid medium using acetate as the

carbon and energy source. Relatives of several of these

species were isolated from environments that are similar

to concrete in salinity or alkalinity. For example,Kocuria

species havebeen found inmarine sediments [33] and salt

evaporation works [34]. Dietzia natronolimnaea strains

have been isolated from moderately saline alkaline lakes

[35], and Nocardioides lentus has been isolated from

alkaline soils [36].

4.3 Microbial community characterized

by sequencing

4.3.1 Identification of a suitable DNA extraction

method

Because ancient bone, like concrete, is dry, hard, and

high in Ca2?, methods developed for extraction of

O Si Ca Al K Na Mg Fe S Mo0

10

20

30

40

50

60

70

Wei

ght %

Fig. 2 SEM–EDS analysis of concrete composition. The

weight percent of each element was quantified by SEM–EDS

in 15–20 fields of view in three separate subsamples from a

single cylinder


DNA from ancient bone were modified for application

to concrete samples [27, 37]. Based on these methods,

the sample is washed with EDTA prior to extraction to

remove some of the divalent cations; later, additional

salts are precipitated using a sodium acetate solution.

Using this protocol, we consistently obtain high-

molecular-weight genomic DNA from concrete

(Fig. 3). This DNA is suitable for downstream

processing by polymerase chain reaction (PCR)

amplification and sequencing. No genomic DNA was

observed after extracting DNA from the EDTA wash

solution, confirming that this step does not result in

appreciable loss of biomass from the samples.

4.3.2 General sequence quality

The 16S rRNA genes of the bacterial communities

were amplified from DNA samples from the top and

inside of the test cylinder (Fig. 1c). Sequences of this

gene are standardly used to assess the taxonomic

diversity of microbes in natural environments, as the

16S rRNA gene is both universally distributed and

highly conserved [28, 38]. Amplicons were sequenced

using Illumina MiSeq technology at the Research and

Testing Laboratories in Lubbock, TX. 25,317 paired-

end reads were obtained from the ‘‘top’’ sample, and

59,086 from the ‘‘inside’’ sample. The majority of

these reads could be joined with their mate-pairs and

passed subsequent quality filtering (Table 3).

4.3.3 Bacteria identified by sequencing

More than 250 bacterial genera were identified in the

sequence data sets (Fig. 4; Supplementary Table 1).

The bacterial community on the top surface was more

diverse than that inside the concrete (Fig. 4), but the

Table 2 Identification of bacteria isolated from concrete

Strain designation Phylum Closest relative in GenBank Colony color Cell shape

IN-02CB Alphaproteobacteria Paracoccus marinus KKL-B9 Green–yellow

IN-02 Gammaproteobacteria Pseudoalteromonas sp. BSi20669 Orange Short rod

IN-01 Actinobacteria Arthrobacter sp. 32c White Short rod

IN-02C Actinobacteria Kocuria marina Green–yellow Coccus

IN-03 Actinobacteria Dietzia cercidiphylli strain X10 Orange Rod

IN-04 Actinobacteria Rhodococcus cercidiphylli strain EB37 Orange–yellow Long rod

IN-06 Actinobacteria Arthrobacter sp. clone EK_Ca790 Pink Rod

IN-06B Actinobacteria Kocuria marina Yellow–green Coccus (pairs)

IN-07 Actinobacteria Arthrobacter agilis strain DSM 20550 Pink Coccus

IN-08 Actinobacteria Nocardioides lentus strain KSL-19 Burnt Orange

IN-09 Actinobacteria Dietzia natronolimnaea strain NF047 Orange–red Rod

CL-01 Actinobacteria Dietzia natronolimnaea strain NF047 Orange Short rod

CL-02 Actinobacteria Uncultured Arthrobacter sp. clone Acro223 Yellow–green Coccus

CL-04 Actinobacteria Arthrobacter sp. 5139 White Coccus (pairs)

CL-05 Actinobacteria Rhodococcus erythropolis strain zzx26 Coccus

All strains were identified by amplification of the 16S rRNA gene with primers 8F and 1492R and Sanger sequencing of the products.

Strains designated ‘‘CL’’ were isolated from top or side surfaces of concrete; strains designated ‘‘IN’’ were isolated from internal

samples

Table 3 16S Illumina amplicon read data from microbial community DNA

Sample No. of raw

read pairs

No. of joined

sequences

Joined sequences

passing quality filter

Top 25,317 23,803 23,105

Inside 59,086 49,123 42,797

The majority of raw reads passed quality filtering


most abundant groups in both were Actinobacteria and

Proteobacteria (Fig. 4). In the inside sample, the most

abundant Actinobacterial genera were the Micrococ-

caceae, which includes the Arthrobacter spp. and

Nocardioidaceae that we isolated. Spore-forming

Firmicutes (Bacillus and Clostridium spp.)

are *2.5 % of the sequences in the inside sample;

in total, Gram-positive bacteria comprise more than

65 % of the species identified in the inside sample

(Fig. 4; Supplementary Table 1). The Alphapro-

teobacteria were the most abundant Proteobacteria,

specifically Rhizobia and Caulobacteria (both com-

mon soil bacteria). Nearly 6 % of sequences could not

be assigned to specific taxa.

In contrast, the Actinobacteria only com-

prised *33 % of the total community on the top of

the cylinder. The Alphaproteobacteria and Betapro-

teobacteria species were much greater fractions of this

community, though many of the Alphaproteobacteria

could not be classified further (Fig. 4). Sequences

related to cyanobacteria and chloroplasts (plant-asso-

ciated organelles, likely coming from pollen in these

samples) were more abundant on the top surface than

inside. Putatively spore-forming Firmicutes were only

small fractions of both communities, but were more

abundant in the inside sample. Deltaproteobacteria

related to Bdellovibrio and Myxococcus, which are

predatory species that consume other bacteria, were

observed only in the top sample (Fig. 4).

4.3.4 Potential effects of environmental conditions

on microbial community

Although concrete is a relatively low-diversity envi-

ronment (in comparison, the average soil microbial

community has between 20,000 and 50,000 species

per gram of soil [39]), we adequately sampled the

diversity present in this environment (Fig. 5). The

composition and limited diversity of the microbial

communities in and on concrete likely reflect the

sources of the bacteria, the chemical composition of

the concrete, and the effect(s) of environmental

conditions. Because environmental inputs to concrete

(via rain, deposition, or fecal contamination) and

weather conditions may change on a daily basis, the

community composition likely also reflects the length

of time that the concrete has been exposed. Similar to

microbial biofilms on rock or monument surfaces,

bacteria in and on concrete must be able to tolerate the

fundamental properties of concrete (high pH, high

salinity, low moisture content) as well as environ-

mental conditions, which may change rapidly [40].

From February 2012 to February 2013, these test

cylinders were on a university rooftop, where they

were exposed to temperatures from -13 to 38 �C and

a total of 95.5 cm of precipitation (data for February

10, 2012 to February 10, 2013 obtained from www.

wunderground.com/history). The diurnal and seasonal

variations likely selected for organisms tolerant to a

wide range of conditions or for organisms capable of

forming dormant cell types, which can withstand

extreme environmental conditions.

Actinobacteria were the most abundant organisms

in the concrete-associated bacterial communities

characterized in this study, and are similarly prevalent

in aerobic surface-associated microbial communities

on rock surfaces [40]. The microbial communities on

monuments exposed to weather, such as a 13th-

century wall painting in Austria [41], and endolithic

and epilithic communities in and on Mayan limestone

monuments [42], are also broadly similar to those on

concrete. In these environments, Actinobacteria and

Proteobacteria are highly represented [41, 43–45].

However, many of those communities are dominated

by photosynthetic Cyanobacteria or algae, and often

have more spore-forming Firmicutes as well. In

M1 M21 2 3 4 5 6 0

300

1000

2500

10000

10171

48502

Fig. 3 Genomic DNA extracted from concrete samples. Lanes

M1 and M2 are molecular-weight standards; molecular weights

of specific bands (in base pairs) are indicated to the left and right

of the photo. Lanes 1–5 are genomic DNA extracted from

concrete; Lane 6 is genomic DNA extracted from 1 g of soil, for

comparison. Lane 0 is a negative control. High-molecular-

weight DNA can clearly be seen in lanes 1–6, as expected for

genomic DNA


http://www.wunderground.com/history

http://www.wunderground.com/history

Micrococcaceae


Nocardiodaceae


Other/unclassified

Oxalobacteraceae




Micrococcaceae







Alphaproteobacteria

Firmicutes

Actinobacteria




Deltaproteobacteria

TM-7Chloroflexi

Oxalobacteraceae

Micrococcaceae






















A

B

Top

Inside

*

*

*

*

*

*

*

*

*


contrast, in both the surface and interior communities

of concrete, Firmicutes and Cyanobacteria are minor

components. Cyanobacteria and chloroplasts com-

prise only *1.5 % of the community on the surface,

and *0.6 % in the interior; Firmicutes are *2.5 %

of the organisms inside this sample, and less than

0.3 % of the organisms on the surface. The organisms

identified in our test cylinders are related to bacteria

that are tolerant of if not adapted to alkaline, saline, or

desert environments. For example, bacteria in the

Geodermatophilaceae group, which are known to

inhabit desert soils and the surfaces and interiors of

rocks [46, 47], comprise 3.5 % of the community on

the concrete surface and *1.2 % inside (Fig. 4;

Supplementary Table 1). Similarly, *1 % of the

bacteria both inside and on our concrete sample were

members of the Skermanella genus, representatives of

which have been isolated from airborne dust [48],

desert soils [49] and heavy-metal-enriched soils [50].

The presence of these organisms suggests that some of

the bacteria in concrete may come from the aggregate

or sand, as well as atmospheric deposition.

Previous analysis of microbially-induced concrete

corrosion has focused on surface-colonizing microbes

on the walls and ceilings of concrete sewers

[10, 14, 51, 52]. We observe very little overlap between

the organisms found in the concrete cylinder samples

described here, which are primarily Actinobacteria, and

the acidophiles and sulfate-reducing bacteria found in

corroding concrete systems [10, 14, 51]. Although one

study found that Actinobacteria and Alphaproteobac-

teria dominated the microbial communities on concrete

sewer surfaces [51], the species distribution was quite

different, as the Actinobacteria in those samples were

primarilyMycobacterium spp. and the Alphaproteobac-

teria were mostly Acidiphilium spp., neither of which

represents more than 0.01 % of the community in

cylinders characterized here (Supplementary Table 1).

The large-scale differences between microbial commu-

nities in and on dry concrete (this study) and those on

intermittently-submerged sewer surfaces suggest that

environmental conditions play an important role in

development of the surface-associated concrete

microbiome.

Because our test samples were kept on a green roof,

they may also have been exposed to avian fecal matter.

However, the sequences from the top surface do not

have any of the major indicator species identified in

gull or goose [53, 54]. We conclude that the inputs

contributing either novel microbes or nutrients to these

concrete samples included precipitation and atmo-

spheric deposition, but likely not fecal material from

birds.

4.3.5 Comparison of bacteria identified

by sequencing and cultivation

The microbial community in the interior of the

concrete sample had much greater actinobacterial

and proteobacterial diversity than we isolated. Over 60

genera of Actinobacteria were detected by amplicon

sequencing (Supplementary Table 1). All of the

actinobacterial isolates belong to genera identified in

the sequence data set. However, the genera repre-

sented in the culture collection are not necessarily the

most numerically abundant in the sequence data set

(Table 2; Fig. 4; our laboratory has isolated represen-

tatives of taxa with asterisks). Although the

Corynebacteria are six of our 13 actinobacterial

bFig. 4 Bacterial taxa present inside (a) or on the surface of (b) aconcrete cylinder, plotted in order of abundance.Colors indicate

the phylum-level classifications; genera are indicated in

parentheses where they could be classified. Groups representing

less than 0.5 % of the community are not labeled, but can be

found in Supplementary Table 1. Asterisks indicate groups that

include species found in our culture collection

4000 8000 12000 1600000

200

400

600

800

1000

1200

1400

sequences per subsample

obse

rved

OTU

s

TopInside

Fig. 5 Rarefaction analysis. The sequence data was subsam-

pled (ten replicates for each level of subsampling) and the

average number of OTUs observed in each subsample plotted.

Error bars indicate one standard deviation. The lines for both

the inside and top sample approach a plateau, indicating that our

sequence coverage was sufficient to sample the majority of the

taxa present in and on concrete


isolates, they represent less than 0.3 % of the

sequences from the ‘‘inside’’ sample, and an even

smaller fraction of the ‘‘top’’ sample (Supplementary

Table 1). However, Micrococcales are well-repre-

sented in both the sequence data sets and in our

collection of isolates. Better understanding how these

isolates survive and grow in concrete will therefore

give us representative information about the bacteria

in concrete. The prevalence of Actinobacteria rather

than the spore-forming Firmicutes was unexpected,

since spores could enable cells to survive the harsh

conditions in concrete. However, Actinobacteria have

similarly resistant membranes and some species form

cysts or other dormant cell types [55–57]. The

carotenoid pigments in Actinobacteria have also been

shown to protect the cells from temperature stress and

salt stress [58], which suggests a possible role for the

pigments in the isolates described here.

Although the organisms that we isolated represent

abundant groups within the concrete microbial com-

munity, they represent only a small fraction of the

diversity present. This discrepancy is common when

comparing cultivation-dependent and cultivation-in-

dependent analyses of the same environment [59, 60].

Additional cultivation efforts using different media

formulations optimized for the kinds of bacteria

observed in this environment would likely result in

isolation of more species, including the numerically

abundant Alphaproteobacteria.

4.3.6 Alpha diversity analysis

To determine whether our sequence coverage was

enough to detect the majority of species present, we

performed rarefaction analyses in which the number of

OTUs observed in subsampled data were quantified

(Fig. 5). Both curves in the rarefaction approach a

plateau, indicating that our sequencing effort was

sufficient to sample the majority of the OTUs in

concrete. This data was then used to the calculate the

Chao1 richness estimator, which uses the number of

observed OTUs in subsampled data to extrapolate the

total number of taxa in a sample [61, 62]. We calculate

that approximately 1300 OTUs are present in both

samples, and observe just over 1100 in both. Addi-

tional sequencing effort could reveal additional rare

taxa, but would not change the overall community

composition described here.

5 Conclusions

The objectives of this study were to confirm the

presence of viable bacteria in concrete, develop a

method for extraction of DNA from hardened con-

crete, and characterize bacterial communities in and

on concrete. Here, we show that viable bacteria are

present in concrete up to a year after pouring and

curing, as demonstrated by successful cultivation and

isolation of 15 different bacterial strains. We further

demonstrate that DNA can be extracted from concrete

and is pure enough to be amplified and sequenced, and

that the microbial community in and on concrete is

dominated by two bacterial phyla, the Actinobacteria

and Proteobacteria. The DNA extraction method can

be used to characterize microbial communities in

different types of concrete, in concrete exposed to

different environmental conditions, or in corroding

concrete.

Acknowledgments This work was supported by grant #

12A01559 from the University of Delaware Research

Foundation to JAM and TS. Additional support was provided

by the Mid-Atlantic Transportation Sustainability University

Transportation Center (MATS UTC). MATS UTC is funded by

grant # DTRT13-G-UTC33 from the US Department of

Transportation and matching funds organized by the Delaware

Center for Transportation. We thank Keira Zhang for assistance

with concrete processing. Additionally, the authors gratefully

acknowledge Dr. Deborah Powell at the University of Delaware

BioImaging Center for assistance with the SEM/EDS analysis,

Dr. Brewster Kingham at the University of Delaware

Sequencing and Genotyping Center for sequencing of the 16S

rRNA genes of the isolates and the Research and Testing

Laboratories in Lubbock, TX, for amplicon sequencing of the

16S rRNA genes from DNA extracted directly from concrete.

We thank David Dodd at the Delaware Department of

Transportation for concrete materials and mix designs, Dr.

Jennifer Biddle and Dr. Farshad Rajabipoor for helpful

discussions, and Gary Wenczel, Michael Davidson, and Dr.

Jessica Keffer for technical assistance.

References

1. Arndt RW, Schumacher T, Algernon D, Kee S-H (2011)

Strategies for maintenance of highway bridges in the US-

with the support of nondestructive testing and structural

health monitoring. Bautechnik 88:793–804

2. Amini K, Kraatz H-B (2014) Recent advances and devel-

opments in monitoring biological agents in water samples.

Rev Environ Sci Bio/Technol 14:23–48. doi:10.1007/

s11157-014-9351-5

3. Marine SC, Pagadala S, Wang F, Pahl DM, Melendez MV,

Kline WL, Oni RA, Walsh CS, Everts KL, Buchanan RL,


http://dx.doi.org/10.1007/s11157-014-9351-5

http://dx.doi.org/10.1007/s11157-014-9351-5

Micallef SA (2015) The growing season, but not the farming

system, is a food safety risk determinant for leafy greens in

the mid-Atlantic region of the United States. Appl Environ

Microbiol 81:2395–2407. doi:10.1128/AEM.00051-15

4. O’Connell M, McNally C, Richardson MG (2010) Bio-

chemical attack on concrete in wastewater applications: a

state of the art review. Cem Concr Compos 32:479–485.

doi:10.1016/j.cemconcomp.2010.05.001

5. Wei S, Sanchez M, Trejo D, Gillis C (2010) Microbial

mediated deterioration of reinforced concrete structures. Int

Biodeterior Biodegrad 64:748–754. doi:10.1016/j.ibiod.

2010.09.001

6. Wei S, Jiang Z, Liu H, Zhou D, Sanchez-Silva M (2013)

Microbiologically induced deterioration of concrete—a

review. Braz J Microbiol 44:1001–1007. doi:10.1590/

S1517-83822014005000006

7. Jonkers HM, Thijssen A, Muyzer G, Copuroglu O, Sch-

langen E (2010) Application of bacteria as self-healing

agent for the development of sustainable concrete. Ecol Eng

36:230–235

8. Van Tittelboom K, De Belie N, De Muynck W, Verstraete

W (2010) Use of bacteria to repair cracks in concrete. Cem

Concr Res 40:157–166

9. De Muynck W, De Belie N, Verstraete W (2010) Microbial

carbonate precipitation in construction materials: a review.

Ecol Eng 36:118–136. doi:10.1016/j.ecoleng.2009.02.006

10. Okabe S, Odagiri M, Ito T, Satoh H (2007) Succession of

sulfur-oxidizing bacteria in the microbial community on

corroding concrete in sewer systems. Appl Environ

Microbiol 73:971–980. doi:10.1128/AEM.02054-06

11. Taylor CB, Hutchinson GH (1947) Corrosion of concrete

caused by sulphur-oxidising bacteria. J Soc Chem Ind

66:54–57. doi:10.1002/jctb.5000660205

12. Trejo D, de Figueiredo P, Sanchez M, Gonzalez C, Shiping

W, Li L (2008) Analysis and assessment of microbial bio-

film-mediated concrete deterioration. Texas Transportation

Institute, College Station

13. Gu J-D, Ford TE, Mitchell R (2011) Microbiological cor-

rosion of concrete. In: Revie RW (ed) Uhlig’s corrosion

handbook, 3rd edn. Wiley, New York, pp 451–460

14. Gomez-Alvarez V, Revetta RP, Santo Domingo JW (2012)

Metagenome analyses of corroded concrete wastewater pipe

biofilms reveal a complex microbial system. BMC Micro-

biol 12:122. doi:10.1186/1471-2180-12-122

15. Ling AL, Robertson CE, Harris JK, Frank DN, Kotter CV,

Stevens MJ, Pace NR, Hernandez MT (2014) Carbon

dioxide and hydrogen sulfide associations with regional

bacterial diversity patterns in microbially induced concrete

corrosion. Environ Sci Technol 48:7357–7364. doi:10.

1021/es500763e

16. Gomez-Alvarez V (2014) Biofilm-growing bacteria

involved in the corrosion of concrete wastewater pipes:

protocols for comparative metagenomic analyses. Methods

Mol Biol 1147:323–340. doi:10.1007/978-1-4939-0467-9_

23

17. Ling AL, Robertson CE, Harris JK, Frank DN, Kotter CV,

Stevens MJ, Pace NR, Hernandez MT (2015) High-resolu-

tion microbial community succession of microbially

induced concrete corrosion in working sanitary manholes.

PLoSOne 10:e0116400. doi:10.1371/journal.pone.0116400

18. De Muynck W, Cox K, De Belie N, Verstraete W (2008)

Bacterial carbonate precipitation as an alternative surface

treatment for concrete. Constr Build Mater 22:875–885.

doi:10.1016/j.conbuildmat.2006.12.011

19. Jonkers HM, Schlangen E (2007) Crack repair by concrete-

immobilized bacteria. In: Proceedings of the first interna-

tional conference on self healing materials

20. Lv J, Mao J, Ba H (2015) Influence of marine microor-

ganisms on the permeability and microstructure of mortar.

Constr Build Mater 77:33–40. doi:10.1016/j.conbuildmat.

2014.11.072

21. Berndt ML (2011) Evaluation of coatings, mortars and mix

design for protection of concrete against sulphur oxidising

bacteria. Constr Build Mater 25:3893–3902. doi:10.1016/j.

conbuildmat.2011.04.014

22. Soleimani S, Ormeci B, Isgor OB (2013) Growth and char-

acterization of Escherichia coli DH5a biofilm on concrete

surfaces as a protective layer against microbiologically

influenced concrete deterioration (MICD). Appl Microbiol

Biotechnol 97:1093–1102. doi:10.1007/s00253-012-4379-3

23. Wolin EA, Wolin MJ, Wolfe RS (1963) Formation of

methane by bacterial extracts. J Biol Chem 238:2882–2886

24. Turner S, Pryer KM, Miao VPW, Palmer JD (1999) Inves-

tigating deep phylogenetic relationships among cyanobac-

teria and plastids by small subunit rRNA sequence analysis.

J Eukaryot Microbiol 46:327–338. doi:10.1111/j.1550-

7408.1999.tb04612.x

25. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z,

Miller W, Lipman DJ (1997) Gapped BLAST and PSI-

BLAST: a new generation of protein database search pro-

grams. Nucleic Acids Res 25:3389–3402

26. Hugenholtz P, Pitulle C, Hershberger KL, Pace NR (1998)

Novel division level bacterial diversity in a Yellowstone hot

spring. J Bacteriol 180:366–376

27. Cattaneo C, Smillie DM, Gelsthorpe K, Piccinini A, Gel-

sthorpe AR, Sokol RJ (1995) A simple method for

extracting DNA from old skeletal material. Forensic Sci Int

74:167–174. doi:10.1016/0379-0738(95)01758-B

28. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D,

Huntley J, Fierer N, Owens SM, Betley J, Fraser L, Bauer

M, Gormley N, Gilbert JA, Smith G, Knight R (2012) Ultra-

high-throughput microbial community analysis on the

Illumina HiSeq andMiSeq platforms. ISME J 6:1621–1624.

doi:10.1038/ismej.2012.8

29. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K,Bushman FD, Costello EK, Fierer N, Gonzalez Pena A,

Goodrich JK, Gordon JI, Huttley GA, Kelley ST, Knights D,

Koenig JE, Ley RE, Lozupone CA, McDonald D, Muegge

BD, Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ,

Walters WA, Widmann J, Yatsunenko T, Zaneveld J,

Knight R (2010) QIIME allows analysis of high-throughput

community sequencing data. Nat Methods 7:335–336.

doi:10.1038/nmeth.f.303

30. Aronesty E (2013) Comparison of sequencing utility pro-

grams. Open Bioinform J 7:1–8

31. Kuczynski J, Stombaugh J, Walters WA, Gonzalez A,

Caporaso JG, Knight R (2012) Using QIIME to analyze 16S

rRNA gene sequences from microbial communities. Curr

Protoc Microbiol Chapter 1:Unit 1E.5. doi: 10.1002/

9780471729259.mc01e05s27


http://dx.doi.org/10.1128/AEM.00051-15

http://dx.doi.org/10.1016/j.cemconcomp.2010.05.001

http://dx.doi.org/10.1016/j.ibiod.2010.09.001


http://dx.doi.org/10.1590/S1517-83822014005000006

http://dx.doi.org/10.1590/S1517-83822014005000006

http://dx.doi.org/10.1016/j.ecoleng.2009.02.006


http://dx.doi.org/10.1002/jctb.5000660205

http://dx.doi.org/10.1186/1471-2180-12-122

http://dx.doi.org/10.1021/es500763e

http://dx.doi.org/10.1021/es500763e

http://dx.doi.org/10.1007/978-1-4939-0467-9_23

http://dx.doi.org/10.1007/978-1-4939-0467-9_23

http://dx.doi.org/10.1371/journal.pone.0116400

http://dx.doi.org/10.1016/j.conbuildmat.2006.12.011





http://dx.doi.org/10.1007/s00253-012-4379-3

http://dx.doi.org/10.1111/j.1550-7408.1999.tb04612.x

http://dx.doi.org/10.1111/j.1550-7408.1999.tb04612.x

http://dx.doi.org/10.1016/0379-0738(95)01758-B

http://dx.doi.org/10.1038/ismej.2012.8

http://dx.doi.org/10.1038/nmeth.f.303

http://dx.doi.org/10.1002/9780471729259.mc01e05s27

http://dx.doi.org/10.1002/9780471729259.mc01e05s27

32. Yamanaka T, Aso I, Togashi S, Tanigawa M, Shoji K,

Watanabe T, Watanabe N, Maki K, Suzuki H (2002) Cor-

rosion by bacteria of concrete in sewerage systems and

inhibitory effects of formates on their growth. Water Res

36:2636–2642. doi:10.1016/S0043-1354(01)00473-0

33. Kim SB, Nedashkovskaya OI, Mikhailov VV, Han SK, Kim

K-O, Rhee M-S, Bae KS (2004) Kocuria marina sp. nov., a

novel actinobacterium isolated from marine sediment. Int J

Syst Evol Microbiol 54:1617–1620. doi:10.1099/ijs.0.

02742-0

34. Sarafin Y, Donio MBS, Velmurugan S, Michaelbabu M,

Citarasu T (2014) Kocuria marina BS-15 a biosurfactant

producing halophilic bacteria isolated from solar salt works

in India. Saudi J Biol Sci 21:511–519. doi:10.1016/j.sjbs.

2014.01.001

35. Duckworth A, Grant S, GrantW, Jones BE, Meijer D (1998)

Dietzia natronolimnaios sp. nov., a new member of the

genus Dietzia isolated from an East African soda lake.

Extremophiles 2:359–366

36. Yoon J-H, Lee C-H, Oh T-K (2006) Nocardioides lentus sp.

nov., isolated from an alkaline soil. Int J Syst Evol Micro-

biol 56:271–275. doi:10.1099/ijs.0.63993-0

37. Loreille OM, Diegoli TM, Irwin JA, Coble MD, Parsons TJ

(2007) High efficiency DNA extraction from bone by total

demineralization. Forensic Sci Int Genet 1:191–195. doi:10.

1016/j.fsigen.2007.02.006

38. Reysenbach AL, Wickham GS, Pace NR (1994) Phyloge-

netic analysis of the hyperthermophilic pink filament com-

munity in Octopus Spring, Yellowstone National Park. Appl

Environ Microbiol 60:2113–2119

39. Roesch LFW, Fulthorpe RR, Riva A, Casella G, Hadwin

AKM, Kent AD, Daroub SH, Camargo FAO, Farmerie WG,

Triplett EW (2007) Pyrosequencing enumerates and con-

trasts soil microbial diversity. ISME J 1:283–290. doi:10.

1038/ismej.2007.53

40. Gorbushina AA (2007) Life on the rocks. EnvironMicrobiol

9:1613–1631. doi:10.1111/j.1462-2920.2007.01301.x

41. Rolleke S, Muyzer G, Wawer C, Wanner G, Lubitz W

(1996) Identification of bacteria in a biodegraded wall

painting by denaturing gradient gel electrophoresis of PCR-

amplified gene fragments coding for 16S rRNA. Appl

Environ Microbiol 62:2059–2065

42. McNamara CJ, Perry TD, Bearce KA, Hernandez-Duque G,

Mitchell R (2006) Epilithic and endolithic bacterial com-

munities in limestone from a Maya archaeological site.

Microb Ecol 51:51–64. doi:10.1007/s00248-005-0200-5

43. Ramırez M, Hernandez-Marine M, Novelo E, Roldan M

(2010) Cyanobacteria-containing biofilms from a Mayan

monument in Palenque, Mexico. Biofouling 26:399–409

44. Suihko M-L, Alakomi H-L, Gorbushina A, Fortune I,

Marquardt J, Saarela M (2007) Characterization of aerobic

bacterial and fungal microbiota on surfaces of historic

Scottish monuments. Syst Appl Microbiol 30:494–508.

doi:10.1016/j.syapm.2007.05.001

45. Gaylarde CC, Gaylarde PM, Beech IB (2008) Deterioration of

limestone structures associated with copper staining. Int Biode-

terior Biodegrad 62:179–185. doi:10.1016/j.ibiod.2008.01.007

46. Normand P, Daffonchio D, Gtari M (2014) The family

Geodermatophilaceae. In: Rosenberg E, DeLong EF, Lory

S, Stackebrandt E, Thompson F (eds) The Prokaryotes.

Springer, Berlin, pp 361–379

47. Urzi C, Brusetti L, Salamone P, Sorlini C, Stackebrandt E,

Daffonchio D (2001) Biodiversity of Geodermatophilaceae

isolated from altered stones and monuments in the

Mediterranean basin. EnvironMicrobiol 3:471–479. doi:10.

1046/j.1462-2920.2001.00217.x

48. Weon H-Y, Kim B-Y, Hong S-B, Joa J-H, Nam S-S, Lee

KH, Kwon S-W (2007) Skermanella aerolata sp. nov., iso-

lated from air, and emended description of the genus

Skermanella. Int J Syst Evol Microbiol 57:1539–1542.

doi:10.1099/ijs.0.64676-0

49. An H, Zhang L, Tang Y, Luo X, Sun T, Li Y,Wang Y, Dai J,

Fang C (2009) Skermanella xinjiangensis sp. nov., isolated

from the desert of Xinjiang, China. Int J Syst EvolMicrobiol

59:1531–1534. doi:10.1099/ijs.0.003350-0

50. Luo G, Shi Z, Wang H, Wang G (2012) Skermanella

stibiiresistens sp. nov., a highly antimony-resistant bac-

terium isolated from coal-mining soil, and emended

description of the genus Skermanella. Int J Syst Evol

Microbiol 62:1271–1276. doi:10.1099/ijs.0.033746-0

51. Cayford BI, Dennis PG, Keller J, Tyson GW, Bond PL

(2012) High-throughput amplicon sequencing reveals dis-

tinct communities within a corroding concrete sewer sys-

tem. Appl Environ Microbiol 78:7160–7162

52. Vupputuri S, Fathepure BZ, Wilber GG, Sudoi E, Nas-

razadani S, LeyMT, Ramsey JD (2015) Isolation of a sulfur-

oxidizing Streptomyces sp. from deteriorating bridge

structures and its role in concrete deterioration. Int Biode-

terior Biodegrad 97:128–134. doi:10.1016/j.ibiod.2014.11.

002

53. Lu J, Santo Domingo JW, Lamendella R, Edge T, Hill S

(2008) Phylogenetic diversity and molecular detection of

bacteria in gull feces. Appl Environ Microbiol

74:3969–3976. doi:10.1128/AEM.00019-08

54. Green HC, Dick LK, Gilpin B, Samadpour M, Field KG

(2012) Genetic markers for rapid PCR-based identification

of gull, Canada goose, duck, and chicken fecal contamina-

tion in water. Appl Environ Microbiol 78:503–510. doi:10.

1128/AEM.05734-1155. Demkina EV, Soina VS, Registan GIE, Zvyagintsev DG

(2000) Reproductive resting forms ofArthrobacter globi-

formis. Microbiology 69:309–313. doi:10.1007/BF0275

6739

56. Shimkets LJ (2013) Prokaryotic life cycles. In: Rosenberg

E, DeLong EF, Lory S, Stackebrandt E, Thompson F (eds)

Prokaryotes prokaryotic communities ecophysiol. Springer,

Berlin, pp 317–336

57. Young M, Artsatbanov V, Beller HR, Chandra G, Chater

KF, Dover LG, Goh E-B, Kahan T, Kaprelyants AS, Kyr-

pides N, Lapidus A, Lowry SR, Lykidis A, Mahillon J,

Markowitz V, Mavromatis K, Mukamolova GV, Oren A,

Rokem JS, Smith MCM, Young DI, Greenblatt CL (2010)

Genome sequence of the Fleming strain of Micrococcus

luteus, a simple free-living actinobacterium. J Bacteriol

192:841–860. doi:10.1128/JB.01254-09

58. Fong NJC, Burgess ML, Barrow KD, Glenn DR (2001)

Carotenoid accumulation in the psychrotrophic bacterium

Arthrobacter agilis in response to thermal and salt stress.

Appl Microbiol Biotechnol 56:750–756. doi:10.1007/

s002530100739

59. Giannantonio DJ, Kurth JC, Kurtis KE, Sobecky PA (2009)

Molecular characterizations of microbial communities


http://dx.doi.org/10.1016/S0043-1354(01)00473-0

http://dx.doi.org/10.1099/ijs.0.02742-0


http://dx.doi.org/10.1016/j.sjbs.2014.01.001

http://dx.doi.org/10.1016/j.sjbs.2014.01.001


http://dx.doi.org/10.1016/j.fsigen.2007.02.006

http://dx.doi.org/10.1016/j.fsigen.2007.02.006



http://dx.doi.org/10.1111/j.1462-2920.2007.01301.x

http://dx.doi.org/10.1007/s00248-005-0200-5

http://dx.doi.org/10.1016/j.syapm.2007.05.001


http://dx.doi.org/10.1046/j.1462-2920.2001.00217.x

http://dx.doi.org/10.1046/j.1462-2920.2001.00217.x









http://dx.doi.org/10.1007/BF02756739

http://dx.doi.org/10.1007/BF02756739

http://dx.doi.org/10.1128/JB.01254-09

http://dx.doi.org/10.1007/s002530100739

http://dx.doi.org/10.1007/s002530100739

fouling painted and unpainted concrete structures. Int

Biodeterior Biodegrad 63:30–40. doi:10.1016/j.ibiod.2008.

06.004

60. Epstein S (2013) The phenomenon of microbial uncul-

tivability. Curr Opin Microbiol 16:636–642. doi:10.1016/j.

mib.2013.08.003

61. Chao A (1984) Nonparametric estimation of the number of

classes in a population. Scand J Stat 11:265–270

62. Colwell RK, Coddington JA (1994) Estimating terrestrial

biodiversity through extrapolation. Philos Trans R Soc Lond

B Biol Sci 345:101–118. doi:10.1098/rstb.1994.0091




http://dx.doi.org/10.1016/j.mib.2013.08.003

http://dx.doi.org/10.1016/j.mib.2013.08.003

http://dx.doi.org/10.1098/rstb.1994.0091

dctr final coverpage maresca - mats utc€¦ · concrete is the most widely used building material...

Documents