Routes to Food Spoilage — Salt Fungi and a New Apple Disease, Paecilomyces Rot

ROUTES TO FOOD SPOILAGE — SALT FUNGI AND A NEW APPLE DISEASE, PAECILOMYCES ROT

A Dissertation Presented to the Faculty of the Graduate School

of Cornell University In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

by Megan Nicole Biango-Daniels May 2018

ROUTES TO FOOD SPOILAGE — SALT FUNGI AND A NEW APPLE DISEASE, PAECILOMYCES ROT

Megan Nicole Biango-Daniels, Ph. D. Cornell University 2018

Extremophilic fungi have surprising powers, surviving and growing under inhospitable conditions including high salinity and high temperatures. These abilities make them formidable enemies in food and agricultural environments. Considering these fungi as a continuum in both environments offers a novel way to examine problems that can arise in the field and end on the shelf. Challenging long-held notions about where spoilage fungi can originate, sea salts were examined as potential sources of low water activity spoilage fungi and were found to contain spoilage-species of

Aspergillus, Penicillium, and Cladosporium. Exploring another group of extreme fungi, the farm-to-food transmission route of the heat-resistant spoilage mold, Paecilomyces niveus (Byssochlamys nivea) was elucidated. It was found in a third of New York orchard soils sampled. Experimental inoculations demonstrated its ability to cause a postharvest disease of apples in accordance with Koch’s postulates. The disease was later demonstrated on developing fruit in apple orchards. The newly described disease, Paecilomyces rot, caused by P. niveus presents with symptoms including circular, concentrically-ringed, brown lesions. Its firm, U-shaped internal rot resembles other fungal apple diseases. Importantly, Paecilomyces niveus was observed to produce its heat-resistant ascospores inside apples. When infected apples were used to make juice concentrate using common industry protocols, with stringent thermal processing, some ascospores survived in the finished product. The results suggest that

when lesser-quality apples, infected with P. niveus, are used to make concentrate they may introduce spoilage inoculum as well as patulin, the most significant apple mycotoxin. To evaluate control methods for this new disease in the field, the efficacy of EC50 concentrations of difenoconazole, fludioxonil, and pyrimethanil, three active ingredients in fungicides, were tested on thirty P. niveus isolates. Cultures from agricultural environments, likely to have been previously exposed to fungicides, were found to be less sensitive than cultures from environments, such as food, where prior exposure was unlikely. Although resistance does not appear to be widespread, the exposed isolates were significantly less sensitive to fludioxonil. Future research should explore the spoilage implications of this apple disease, its incidence and etiology, and its control before and after harvest.

BIOGRAPHICAL SKETCH

A native of the Ithaca area, Megan N. Biango-Daniels received her B.S. in Biology with a concentration in Ecology, Evolution, and Behavior from Binghamton University before coming to Cornell.

In her free time, she likes to write haikus.

Cold rain, hot coffee. Warm autoclave hisses “Good morning” heat and pressure waiting.

Plastic pipette tips, their clear, blue, or yellow tips filtered and empty.

Microscope light glows, glass coverslip pressed over. Clipped in place on stage.

To Christopher & To all the people who inspire me but will never read this.

ACKNOWLEDGMENTS

I am grateful for the support of all my family and friends. I would like to thank my advisor Kathie Hodge for showing me what it means to be an excellent scientist, teacher, and mentor. I am grateful to my entire committee: Randy Worobo, Eric

Nelson, and Kerik Cox for their constant encouragement, advice, and support.

I would like to thank the Royall T. Moore fund and Plant Pathology and Plant-microbe

Biology department for supporting my work. I love this department and consider myself privileged to have worked alongside its faculty, who constantly inspire me.

One person, who I want to recognize for always making time to talk with and mentor me is George Hudler; reading his book inspired me to study fungi and I never imagined that I’d be lucky enough to pick his brain.

I am grateful to Cornell Cooperative Extension agents, apple consultants, and orchard owners for their help in soil sampling. I would like to thank Cornell University

Orchards for their support of this work, and Erika Mudrak of the Cornell Statistical

Consulting Unit for her assistance with data analysis.

My work was supported by the USDA National Institute of Food and Agriculture,

Hatch project 1002546, and by the Cornell CALS Arthur Boller Apple Research

Grant. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the

National Institute of Food and Agriculture (NIFA) or the United States Department of

vii

Agriculture (USDA).

viii

TABLE OF CONTENTS

BIOGRAPHICAL SKETCH v

DEDICATION vi

ACKNOWLEDGMENTS vii

LIST OF TABLES xi

LIST OF FIGURES xii

LIST OF ILLUSTRATIONS xiv

Chapter 1: SEA SALTS AS A POTENTIAL SOURCE OF FOOD SPOILAGE FUNGI

INTRODUCTION 1 MATERIALS AND METHODS 2 RESULTS 7 DISCUSSION 15 ACKNOWLEDGMENTS 19 REFERENCES 20

CHAPTER 2: PAECILOMYCES ROT: A NEW APPLE DISEASE

INTRODUCTION 28 MATERIALS AND METHODS 30 RESULTS 42 DISCUSSION 50 ACKNOWLEDGMENTS 55 REFERENCES 56

CHAPTER 3: SURVIVAL OF PAECILOMYCES NIVEUS IN APPLE JUICE CONCENTRATE MADE FROM FRUIT WITH PAECILOMYCES ROT

INTRODUCTION 63 MATERIALS AND METHODS 67 RESULTS 73 DISCUSSION 75 ACKNOWLEDGMENTS 78 REFERENCES 79

CHAPTER 4: PAECILOMYCES NIVEUS: PATHOGENICITY IN THE ORCHARD AND SENSITIVITY TO THREE FUNGICIDES

INTRODUCTION 86 MATERIALS AND METHODS 89 RESULTS 97 DISCUSSION 103 ACKNOWLEDGMENTS 107 REFERENCES 108

LIST OF TABLES

Table 1.1: Identities of fungi isolated from seven commercially available 8 salts.

Table 2.1: Isolates and accession numbers used in phylogenetic analyses. 33

Table 2.2: Origin information for P. niveus isolates from soil in New York 45 orchards.

Table 4.1: Origin information for exposed and baseline P. niveus isolates 90 used to characterize fungicide sensitivity with the NCBI accession numbers for their BenA region sequence.

LIST OF FIGURES

Figure 1.1: The number of fungi in each of the seven salts, organized by 13 ocean of origin.

Figure 1.2: The total number of fungal isolates from the combined 14 subsamples of each salt, with colors representing the different taxa.

Figure 1.3: The global distribution of fungi across subsamples grouped by 15 ocean of origin of salts (Atlantic or Pacific).

Figure 2.1: Phylogenetic tree of concatenated partial β-tubulin and ITS 43 sequences of 31 taxa, including Paecilomyces niveus (Byssochlamys nivea) isolate CO7, rooted with Thermoascus crustaceus as an outgroup.

Figure 2.2: Phylogenetic tree of partial β-tubulin sequences of 47 taxa, 44 including Paecilomyces niveus (Byssochlamys nivea) isolates from New York orchards, rooted with Thermoascus crustaceus as an outgroup.

Figure 2.3: Increase in the mean diameter of lesions inoculated with 50 Paecilomyces niveus (Byssochlamys nivea) compared to the mean diameter of mock-inoculated control wounds on ‘Gala’ and ‘Golden Delicious’ apples.

Figure 3.1: The mean colony forming units per liter of P. niveus from 74 samples taken throughout the apple juice concentrate process.

Figure 3.2: Representative images of P. niveus spore and the concentration 74 of ascospores observed throughout the apple juice concentrate process.

Figure 3.3: The patulin concentration in single strength cider made from 75 mock-inoculated apples and throughout the apple juice concentrate process of concentrate made from apples with Paecilomyces rot.

Figure 4.1: Phylogenetic relationships of P. niveus isolates from this study 98 to related species, rooted with Thermoascus crustaceus as an outgroup.

Figure 4.2: The diameter of inoculated and mock-inoculated lesions on 99 orchard inoculated ‘Gala’ apples at harvest.

Figure 4.3: Boxplot showing the median percent relative growth after one 100 week of baseline and exposed P. niveus isolates at adjusted mean EC50 concentrations of fungicides.

xii

Figure 4.4: Heatmap of relationships among exposed and baseline P. 102 niveus isolates based on percent relative growth at EC50 concentrations of difenoconazole, fludioxonil, and pyrimethanil.

xiii

LIST OF ILLUSTRATIONS

Illustration 2.1: Representative images of post-harvest disease symptoms 47 caused by Paecilomyces niveus (Byssochlamys nivea) on ‘Gala’ and ‘Golden Delicious’ apples.

Illustration 2.2: Representative image of Paecilomyces niveus 48 (Byssochlamys nivea) asci containing ascospores, observed in the cores of infected apples.

Illustration 3.1: Representative images of a ‘Gala’ apple with post-harvest 68 Paecilomyces rot.

Illustration 3.2: Diagram of the apple juice concentrate process used to 69 make concentrate from apples with Paecilomyces rot.

Illustration 4.1: Representative image of an inoculated ‘Gala’ apple, 94 showing how roundness was measured.

Illustration 4.2: Representative image of symptoms of Paecilomyces Rot on 99 ‘Gala’ inoculated in the orchard.

Illustration 4.3: Representative images of the mesocarp of inoculated ‘Gala’ 100 apples showing unique symptoms of Paecilomyces rot in the orchard.

xiv CHAPTER 1

SEA SALTS AS A POTENTIAL SOURCE OF FOOD SPOILAGE FUNGI1

INTRODUCTION

Valued for their nuanced taste and texture and their often artisanal origins, sea salts have lately seen a surge in popularity (Parks, 2014). Salt is widely perceived to be a chemically pure and sterile food ingredient, however, sea salts may carry microbial contaminants (Butinar et al., 2011). Sea salt was shown to be the source of a mycotoxin-producing mold that spoiled dry-cured meat in a Slovenian production facility (Sonjak et al., 2011a). This finding, and the ways in which sea salts are produced and handled, raised the question of the presence of spoilage fungi in sea salts.

One reason sea salts may contain viable fungi is the time-honored method by which salt is harvested from seawater by slow evaporation in shallow ponds called salterns. A third of all salt for human consumption is produced this way (Butinar et al.

2011). Some organisms in saltern ecosystems may persist as viable propagules even when water activity becomes prohibitive to growth. Thus, sea salt may be enriched in microbes that can grow in high salt environments – the very conditions relied on as an important principle of food preservation (Butinar et al., 2011; Gostincar et al., 2010;

1 Biango-Daniels, M.N., and Hodge, K.T. 2018. Sea salts as a potential source of food spoilage fungi. J. Food Microbiology. DOI: 10.1016/j.fm.2017.07.020

1 Grant, 2004; Gunde-Cimerman et al., 2009, 2000; Gunde-Cimerman and Zalar, 2014a;

Javor, 2002; Jay et al., 2008; Zajc et al., 2014).

This study aimed to quantify viable filamentous fungi in sea salts that could cause spoilage when sea salts are used as food ingredients. A selective medium was used to capture living, fast-growing, moderately xerophilic fungi in commercial sea salts. The fungi were identified by DNA sequencing, and fungal communities were examined relative to their ocean of origin.

MATERIALS AND METHODS

Sampling.

This study aimed to assess the fast growing, moderately xerophilic filamentous fungi present in sea salt, which are likely to spoil a range of salted products. Seven salts were sampled, representing a variety of commercially available “Sea Salts” in the

United States. Duplicate packages were purchased of three Atlantic Ocean salts from

France (Fr1, Fr2, and Fr3), three Pacific Ocean salts (Us1, Us2 and Haw), and one mined salt from the Himalayan Mountains (Him). While six of the salts were made from ocean waters, the seventh (Him) was labeled as a sea salt when in fact it was mined from an ancient sea salt deposit: this distinction is noted throughout our analysis. Each duplicate package (true replicates) was sampled two or three separate times, for a total of five subsamples (pseudoreplicates) per type of salt. The salts varied in their appearance and textures. The Atlantic salts (Fr1, Fr2, and Fr3) were fine-grained white or grey salts with no additives. Two of the Pacific salts (Us1 and

2 Us2) were white, fine-grained finishing salts with the additives calcium silicate, dextrose, potassium iodide in both and sodium bicarbonate in one (Us2). The third

Pacific salt (Haw) was a black, medium-coarse salt with no additives. The one mined salt (Him) included was a very coarse, dark pink salt.

Isolation.

Fungi were isolated from each of the 35 subsamples of sea salts (five per each of the seven types of salt). Each subsample was weighed and dissolved in 100ml of room temperature sterile deionized water and immediately filtered through disposable

MicroFunnel (Pall) filter funnels with Supor membranes (0.2µm pore size), all under aseptic conditions. Sterile deionized water without added salt was used as a negative control. The filtered particles, including fungi, were trapped in the membranes, which were removed and placed filtrate-side-up on dichloran-glycerol (DG18) enumeration medium supplemented with 30µg ml -1 streptomycin sulfate (Butinar et al., 2011;

Gunde-Cimerman et al., 2009, 2003, 2000; Hocking and Pitt, 1980). This medium was developed to isolate moderately xerophilic food spoilage fungi. Plates were incubated at room temperature, 24°C (± 2°), for one week. Although the short incubation did not allow for the detection of slow-growing xerophiles, it was ideal to detect fast-growing fungi more likely to induce spoilage concerns. Fungal colonies that emerged in seven days were counted; all were removed aseptically and subcultured on DG18 to obtain unique individual isolates, determined by colony morphology and microscopic characteristics.

3 DNA extraction and PCR amplifications.

Mycelia harvested from DG18 plates were physically disrupted using a pestle in a 1.5 mL microcentrifuge tube containing sterile 0.5mm glass beads (Biospec

Products). Genomic DNA was extracted using the PrepMan Ultra Sample Preparation

Reagent kit (Applied Biosystems by Life Technologies), E.N.Z.A. SP Fungal DNA

Mini Kit (Omega Bio-Tek), or DNeasy Plant Mini Kit (Qiagen) following recommended protocols. Extracted genomic DNA was stored in elution buffer at -

20°C.

DNA from individual isolates was the target of PCR amplification of the internal transcribed spacer (ITS) regions of the rDNA repeat, a widely accepted barcode for fungi (Schoch et al., 2012). The forward ITS5 primer (5’-

GGAGTAAAAGTCGTAACAAGG -3’) and reverse ITS4 primer (5’-

TCCTCCGCTTATTGATATGC-3’) (White et al., 1990) were used in 25µL reactions with Quanta AccuStat Taq DNA polymerase or Sigma-Aldrich REDTaq genomic

DNA polymerase with MgCl2. A DNA Engine PTC-200 thermal cycler (MJ Research) was programmed as follows for the REDTaq PCR: initial denaturation 95°C/2 min, 26 cycles of 94°C/1 min, 55°C/1min, and 72°C/3 min, and final extension of 72°C/15 min. For the AccuStat Taq PCR: initial denaturation 95°C/2 min, 20 cycles of 94°C/1 min, 55°C/30 s, and 72°C/1 min, 15 cycles of 94°C/30 s, 60°C/30 s, and 72°C/1 min, and final extension of 72°C/10 min.

4 Isolate identification.

PCR amplicons were cleaned up using E.N.Z.A. Cycle-Pure (Omega Bio-tek) according to manufacturer instructions and sequenced with a PCR primer using an

Applied Biosystems Automated 3730xl DNA analyzer at the Cornell University

Biotechnology Resource Center. Isolate sequences were identified with NCBI nucleotide BLAST (Agostino, 2012; Altschul et al., 1990). Nucleotide sequences were first blasted (Megablast) against all fungal (taxid: 4751) sequences, excluding uncultured/environmental sample sequences. Because of uncertainty about reputability of the identities of some NCBI accessions, Aspergillus and Penicillium sequences were identified by blasting against a “gold-standard” ITS barcode database assembled from GenBank accession numbers selected by the International Commission on

Penicillium and Aspergillus (Samson et al., 2014; Visagie et al., 2014b).

Cladosporium isolate identities were inferred to the level of species complex by blasting against taxonomically validated sequences from a recent taxonomic monograph (Bensch et al., 2012). Isolate identities were inferred from among BLAST results with a zero E (expect) value, indicating a significant match, by selecting the match with the highest max score, which accounts for quality and length of the alignment (Agostino, 2012). Isolates that matched multiple species equally well were analyzed with the MOLE-BLAST neighbor-joining tree-building tool, a multiple alignment program based on MUSCLE (Edgar, 2004; Saitou and Nei, 1987).

5 Statistical analysis.

To determine whether the quantity of fungal isolates in a sample was associated with salt type, a generalized linear mixed model (GLMM), fitted with a

Poisson model, which is appropriate for discrete, positive responses such as counts, was used (O’Hara, 2009). The GLMM included a fixed effect of salt type and random effect of salt package, as there were multiple subsamples from each of two packages of every salt type. Weight was included as an offset in the model because subsample weights were not identical, ranging from 2.9–5.4g with a mean weight of 3.27g, and more isolates would be expected from larger subsamples. The data were found to be overdispersed (greater observed variance than expected by the model), so a random effect at the observation level was included (Lenth and Hervé, 2015). The effect of salt type on quantity of fungi was assessed with a likelihood ratio test, which was compared to an intercept-only model. These models were executed using the lme4 package in R statistical programming software (Bates et al., 2014; R Core Team,

2015). A post-hoc pairwise comparison of mean CFU/g fungi among the seven salts was performed. Tukey’s adjustment was also performed to avoid false positives caused by multiple comparison issues (Lenth and Hervé, 2015).

To compare the fungal community composition among the seven salts and between salts from Atlantic and Pacific oceans, excluding the mined salt, ordination was done on the six species that were found in more than one subsample. Ordination is a dimension reduction technique that allows visualization of salt communities’ similarities by their proximity to each other in a two-dimensional ordination space

(Ramette, 2007). Here, nonmetric multidimensional scaling (NMDS) was used with

6 the Bray-Curtis dissimilarity index (Ramette, 2007). Species isolated from only one of the salts sampled (including all the Penicillium species) were not included, as they cannot contribute to assessments of similarities among subsamples. Subsamples that contained no species or contained only unique species were excluded, leaving 16 subsamples that were analyzed. Before analysis the data were relativized by subsample total so that each species’ value was its percent abundance in the subsample. Ordination was executed using the Vegan package in R (Oksanen et al.,

2015). The ANOSIM function was used to test for significant differences in communities of different salt types (Ramette, 2007).

RESULTS

Quantification of viable fungi in sea salts.

Of the thirty-five subsamples of sea salts sampled, thirty-two of them contained viable fungi, and there were fungi present in each sea salt. A total of eighty- five fungal isolates were obtained among the subsamples in this study (Table 1.1). The mean number of viable fungal propagules in each salt, measured in colony forming units per gram (CFU/g), ranged from 0.07 to 1.71 (Figure 1.1). Looking at the ocean of origin for the salts, there was no clear difference in number of viable fungi (Figure

1.1). The salts with the fewest propagules were the mined salt, Him, and Us1, with

0.066 and 0.126 CFU/g respectively, while the salt with the greatest CFU/g was Fr2

(Figure 1.1). There was significant variation among salt types in number of fungi

(Figure 1.1), which did not show a package effect (likelihood ratio test =14.875,

7 P=0.021). The total number of isolates from the combined subsamples was highest in

Fr2, which yielded 26 isolates (Figure 1.2).

Table 1.1. Viable fungi isolated from seven commercially available salts and corresponding NCBI accession number. The isolates were deposited in the Food Fungi

Collection at the New York State Agricultural Experiment Station as frozen cultures.

Isolate Best hit (BLAST Ocean Salt NCBI Accession number match) Aspergillus 1100 chevalieri KT826690 (EF652068) Aspergillus 1052 fumigatus KT826642 (EF669931) Aspergillus 1030 sydowii KT826618 (EF652450) Aspergillus 1031 xerophilus KT826619 Atlantic Fr1 (EF652085) Cladosporium 1060, cladosporioides KT826651, KT826652 1065 complex (FJ936159) Cladosporium cladosporioides 1077 KT826669 complex (HM148003) Penicillium 1073 cremeogriseum KT826665 (GU981586) Aspergillus niger 1043 KT826633 (EF661186) 1039A, 1040, Aspergillus KT826628, KT826630, 1049, sydowii KT826639, KT826643, 1054A, (EF652450) KT826687 1097 1017, Atlantic Fr2 1025, 1026, KT826608, KT826613, Aspergillus 1039B, KT826614, KT826629, welwitschiae 1041, KT826631, KT826632, (FJ629340) 1042, KT826634, KT826644 1044, 1054B Aspergillus unguis 1038 KT826627 (EF652443)

9 Table 1.1 Continued KT826649, Cladosporium KT826650, 1058, 1059, cladosporioides KT826653, 1062A, 1066, complex KT826658, 1074, 1083 (FJ936159) KT826666, KT826674 Cladosporium cladosporioides 1088 KT826678 Atlantic Fr2 complex (HM147999) Cladosporium KT826654, 1062B, 1067, herbarum KT826659, 1068 complex KT826660 (EF679334) Ulocladium 1024 dauci KT826612 (NR_077186 ) Aspergillus 1029 niger KT826617 (EF661186) Aspergillus 1069 tubingensis KT826661 (EF661193) KT826609, KT826610, KT826615, 1018, 1019, KT826616, 1027, 1028, Atlantic Fr3 Aspergillus KT826635, 1045, 1046, welwitschiae KT826636, 1047, 1048, (FJ629340) KT826637, 1050, 1051, KT826638, 1052 KT826640, KT826641, KT826642 Cladosporium herbarum 1061 KT826652 complex (EF679334) Aspergillus KT826663, 1071, 1079 caespitosus KT826670 (EF652428) Pacific Haw Aspergillus 1055 fumigatus KT826645 (EF669931)

10 Table 1.1 Continued Calostilbe Pacific Haw 1075 striispora KT826667 (KM231855) Cladosporium KT826606, 1015, 1016, cladosporioides KT826607, 1032 complex KT826620 (HM148003) Cladosporium 1033A, cladosporioides KT826621, 1056A complex KT826646 (FJ936159) Cladosporium cladosporioides 1056B KT826647 complex (HM147994) Cladosporium cladosporioides 1072, 1080 KT826664 Pacific Haw complex (FJ936159) Cladosporium KT826648, 1057, 1064, sphaerospermum KT826656, 1070 complex KT826662 (DQ780364) KT826611, KT826622, 1023, 1033B, Penicillium KT826623, 1034, 1035, oxalicum KT826624, 1036, 1037 (AF033438) KT826625, KT826626 Scopulariopsis 1076 sphaerospora KT826668 (LM652438) Aspergillus KT826685, Pacific Us1 1095, 1096 welwitschiae KT826686 (FJ629340) Cladosporium KT826675, 1084, 1091, cladosporioides KT826681, 1093 complex KT826683 (FJ936159) Pacific Us2 Cladosporium cladosporioides 1089 KT826682 complex (HM147994)

11 Table 1.1 Continued Cladosporium cladosporioides 1092 KT826682 complex (HM147999) Cladosporium cladosporioides 1094 KT826684 complex (HM147998) Cladosporium KT826672, Pacific Us2 1081, 1082, herbarum KT826673, 1087 complex KT826677 (EF679334) Cladosporium herbarum 1090 KT826680 complex (JN906977) Septoriella KT826688, 1098, 1099 phragmitis KT826689 (KR873251) Penicillium Mined Him 1086 camemberti KT826676 (AB479314)

Figure 1.1. Mean viable fungal propagules (C.F.U./g) with standard error for each of the seven salts, organized by ocean of origin. Salts that have the same letter are not significantly different from each other (α=0.05) when no corrections were made for multiple hypotheses tests. After using Tukey’s method for correction, only Fr2 and Us1 were significantly different (α = 0.05).

Fungi in sea salts.

Species composition varied among sea salts but salts from different oceans were united by shared species mainly in the genera Aspergillus and Cladosporium

(Figure 1.2). The most abundant genera were Aspergillus, Cladosporium, and Penicillium, which each contributed multiple species. No Penicillium species was common among two or more salts (Figure 1.2). Species composition varied among salts but several Aspergillus and Cladosporium species were found in multiple salts. In contrast to those shared taxa, 13 taxa were found in low abundance or in only one salt

(Figure 1.2). Sixteen fungal species and three different Cladosporium taxa, representing seven genera, were found among the seven salts (Table 1.1). Mined salt,

13 Him, had no fungi in common with any of the sea salts and yielded only one isolate of

Penicillium camemberti (Figure 1.2).

The six taxa found in multiple sea salts were: Aspergillus fumigatus, A. niger,

A. sydowii, and A. welwitschiae, and the Cladosporium cladosporiodes and C. herbarum complexes (Figure 1.2). Ordination analysis of taxa from different salts found no significant differences among fungal communities (ANOSIM R=0.066,

P=0.24). Similarly, a broader analysis based on ocean of origin also found no significant differences (ANOSIM R=-0.012, P=0.46) (Figure 1.3).

Figure 1.2. Total number of isolates from the combined subsamples of each salt ranged from 1 in mined Him salt to 26 in Fr2. Fungi that were found in only one salt are indicated by * in the legend. Colors correspond to taxa found in each salt, displayed left to right by ocean of origin, with the mined salt, Him, on the far right.

Figure 1.3. Ordination analysis of the global distribution of fungi across salt subsamples grouped by ocean of origin, excluding mined Him salt. Overlapping ellipses, generated by the R package Vegan, represent the standard deviation of NMDS coordinates of samples grouped by ocean of origin and illustrate that fungal community compositions between Atlantic and Pacific salts were not significantly different (P=0.46).

DISCUSSION

Viable fungi in sea salts.

Commercial sea salts consistently contained low levels of a variety of living fungi. This study probably underestimated the true abundance of fungi in sea salt, since only one growth medium was used and very slow-growing fungi were excluded by the one-week incubation time. Fungi were identified using ITS barcoding, which generally works well, although for a few groups species complexes are difficult to resolve (Schoch et al., 2012). Some of the detected fungi have the potential to cause food spoilage when sea salt is used as a food ingredient, especially in products that do

15 not receive heat treatment, as is the case in dry-cured meats in which sea salts were shown to be the source of spoilage mold inoculum (Sonjak et al. 2011a). The origins of these sea salt fungi are unclear— they might arise during production directly from the saltern environment, or by introduction of fungi during processing, handling, or packing. The later source is likely where the single fungal isolates in the mined salt originated.

Sea salts are produced from seawater trapped in solar salterns: flow-through networks of exposed, shallow ponds (Javor 2002). As sea water is progressively concentrated through evaporation, the increasingly saline waters harbor distinctive communities of microbes including bacteria, archaea, algae and fungi (Food and Drug

Administration, 2012; Javor, 2002). These xerophilic species of endemic saltern communities can be carried over into finished salt (Butinar et al., 2005), where they may even contribute to the distinctive flavor, texture, and terroir of sea salts. Fungal propagules that can survive contact with salt may persist in the product for long periods, posing a potential downstream food spoilage risk if they can grow in moderately low water activities common for most foods.

A global saltern community of fungi.

Statistical analysis supports the idea of a pan-global saltern community, in that there was no evidence that sea salt communities are distinct to the Atlantic vs. Pacific oceans (Figure 1.3). Even salts from very distant locations were not significantly different in their fungal composition. Species of Penicillium, Cladosporium, and

Aspergillus predominate in the sea salts examined, and also in salterns (Gunde-

16 Cimerman and Zalar, 2014b). The tested salts contained aspergilli already reported from global saltern communities, including A. niger, A. sydowii, A. chevalieri, A. tubingensis, and A. fumigatus (Figure. 2) (Butinar et al., 2011). Aspergillus nidulans and A. chevalieri are known in other low water activity environments, including foods and salterns (Gesheva and Negoita, 2011; Vytrasová et al., 2002).

Most aspergilli found in salt belong to Aspergillus section Nigri, a group of dark-spored species known for their global saltern distribution and ability to spoil low water activity foods (Butinar et al., 2011; Pitt and Hocking, 2009). Due to their cosmopolitan nature it is impossible to determine whether all the detected isolates originated in salterns or were airborne contaminants (Butinar et al., 2005; Visagie et al., 2014a). Aspergillus welwitschiae was abundant in several salts and may be a previously unidentified saltern fungus. This is the first report of Cladosporium species from finished sea salt. Cladosporium is a prolific, cosmopolitan genus associated with solar salterns and other low water activity environments, as well as indoor and outdoor air (Butinar et al., 2005; Cantrell et al., 2013; Gunde-Cimerman et al., 2003, 2000;

Schubert et al., 2007; Zalar et al., 2007).

Based on their plant pathogenic ecology, Calostilbe striispora and Septoriella phragmitis were likely introduced from plant populations adjacent to salterns or salt storage facilities (Crous et al., 2015; Rossman et al., 1999). Saprobes such as

Ulocladium dauci and Aspergillus caespitosus, known from soil and decay communities, were likely introduced from soil contact during harvest (Runa et al.,

2009; Samson and Mouchacca, 1975). Many penicillia, including P. oxalicum, are recognized as locally abundant saltern species that dominate their niche (Butinar et al.,

17 2011) (Figure 1.2). Penicillium cremeogriseum is unknown from salterns but recognized as an indoor air contaminant (Ejdys and Biedunkiewicz, 2011). It may have been introduced during salt storage after harvest. Him salt, the only mined type of salt in this study, yielded only one isolate which must have been introduced during processing or storage (Figure 1.2).

Salt fungi as spoilage risks.

Fungi detected in sea salts in this study have the significant potential to spoil foods, and introduce mycotoxins or allergens when sea salt is used as a food ingredient

(Bennett and Klich, 2003; Butinar et al., 2011; Frisvad et al., 2004; Ponsone et al.,

2007; Steyn, 1969). Penicillium oxalicum and Scopulariopsis sphaerospora, for example, are important spoilers of sausage and other low water activity foods, and may be introduced to food through salt (Canel et al., 2013; Iacumin et al., 2009;

Papagianni et al., 2007; Wolfe et al., 2014). Penicillium camemberti was isolated from the mined salt and is used to ripen cheeses, but some strains that may be introduced via sea salt produce mycotoxins (Frisvad et al., 2004; Ropars et al., 2012) and so are undesirable in cheese production. Some Cladosporium spp. cause meat spoilage problems in factories and might also originate from sea salts used as ingredients

(Hammami et al., 2014; Mandeel, 2005; Sonjak et al., 2011b; Sørensen et al., 2008).

Multiple species detected in this study, including A. nidulans, A. fumigatus, A. caespitosus, and A. niger, and P. camemberti are capable of producing mycotoxins

(Bräse et al., 2009).

18 All sea salts examined in this small study contained living fungi. This is concerning given that there are no microbiological standards of quality for salt according to the Compendium of Methods for the Microbiological Examinations of

Foods and the Codex Alimentarius of Food and Agricultural Organization (Downes and Ito, 2001; Alimentarius, C., 1985). Standards are needed, as are methods to reduce the risks of sea salts as spoilage mold sources, which might include simple modifications to production and handling practices, such as reducing soil contact and improving storage conditions. Future work will elucidate the full range of fungi in diverse sea salts, and investigate their origins and mitigation in sea salt production.

ACKNOWLEDGMENTS

We are grateful to Erika Mudrak of the Cornell Statistical Consulting Unit for her data analysis advice and expertise, to Katie Wilkins for her assistance creating databases used for identification of Aspergillus, Penicillium, and Cladosporium isolates. We thank Dr. Randy Worobo and Dr. Eric Nelson of Cornell University, and anonymous reviewers for their comments on this manuscript. This work was supported by the

USDA National Institute of Food and Agriculture, Hatch project 1002546. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the National Institute of

Food and Agriculture (NIFA) or the United States Department of Agriculture

(USDA).

19 REFERENCES

Agostino, M., 2012. Introduction to the BLAST suite and BLASTN, in: Practical

Bioinformatics. Garland Science, pp. 47–71.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990. Basic local

alignment search tool. J. Mol. Biol. 215, 403–410. doi:10.1016/S0022-

2836(05)80360-2

Bates, D., Maechler, M., Bolker, B., Walker, S., 2014. lme4: Linear mixed-effects

models using Eigen and S4_R package version 1.1-7 [WWW Document]. URL

http://cran.r-project.org/package=lme4

Bennett, J.W., Klich, M., 2003. Mycotoxins. Clin. Microbiol. Rev. 16, 497–516.

doi:10.1128/CMR.16.3.497

Bensch, K., Braun, U., Groenewald, J.Z., Crous, P.W., 2012. The genus

Cladosporium. Stud. Mycol. 72, 1–401.

doi:10.3114/sim0003.INTRODUCTION

Bräse, S., Encinas, A., Keck, J., Nising, C.F., 2009. Chemistry and biology of

mycotoxins and related fungal metabolites. Chem. Rev. 109, 3903–3990.

doi:10.1021/cr050001f

Butinar, L., Frisvad, J.C., Gunde-Cimerman, N., 2011. Hypersaline waters—a

potential source of foodborne toxigenic Aspergilli and Penicillia. FEMS

Microbiol. Ecol. 77, 186–99. doi:10.1111/j.1574-6941.2011.01108.x

Butinar, L., Zalar, P., Frisvad, J.C., Gunde-Cimerman, N., 2005. The genus

Eurotium—members of indigenous fungal community in hypersaline waters of

20 salterns. FEMS Microbiol. Ecol. 51, 155–166.

doi:10.1016/j.femsec.2004.08.002

Canel, R.S., Wagner, J.R., Stenglein, S. a, Ludemann, V., 2013. Indigenous

filamentous fungi on the surface of Argentinean dry fermented sausages

produced in Colonia Caroya (Córdoba). Int. J. Food Microbiol. 164, 81–6.

doi:10.1016/j.ijfoodmicro.2013.03.022

Cantrell, S. A, Tkavc, R., Gunde-Cimerman, N., Zalar, P., Acevedo, M., Báez-Félix,

C., 2013. Fungal communities of young and mature hypersaline microbial

mats. Mycologia 105, 827–36. doi:10.3852/12-288

Crous, P.W., Carris, L.M., Giraldo, A., Groenewald, J.Z., Hawksworth, D.L.,

Jaklitsch, W.M., Lebrun, M., Schumacher, R.K., Stielow, J.B., Van, E.J., 2015.

The Genera of Fungi - fixing the application of the type species of generic

names—G 2: Allantophomopsis, Latorua, Macrodiplodiopsis, Macrohilum,

Milospium, Protostegia, Pyricularia, Robillarda, Rotula, Septoriella, Torula,

and Wojnowicia. IMA Fungus 6, 163–198.

doi:10.5598/imafungus.2015.06.01.11

Downes, F.P., Ito, K. (Eds.), 2001. Compendium of methods for the microbiological

examination of foods, 4th ed. American Public Heath Association.

Edgar, R.C., 2004. MUSCLE: Multiple sequence alignment with high accuracy and

high throughput. Nucleic Acids Res. 32, 1792–1797. doi:10.1093/nar/gkh340

Ejdys, E., Biedunkiewicz, A., 2011. Fungi of the genus Penicillium in school

buildings. Polish J. Environ. Stud. 20, 333–338.

Food and Drug Administration, 2012. Bad Bug Book: Foodborne pathogenic

21 microorganisms and natural toxins handbook [WWW Document].

http://www.fda.gov/downloads/Food/FoodborneIllnessContaminants/UCM297

627.pdf.

Frisvad, J.C., Smedsgaard, J., Larsen, T.O., Samson, R. A., 2004. Mycotoxins, drugs

and other extrolites produced by species in Penicillium subgenus Penicillium.

Stud. Mycol. 49, 201–241.

Gesheva, V., Negoita, T., 2011. Psychrotrophic microorganism communities in soils

of Haswell Island, Antarctica, and their biosynthetic potential. Polar Biol. 35,

291–297. doi:10.1007/s00300-011-1052-8

Gostincar, C., Grube, M., de Hoog, S., Zalar, P., Gunde-Cimerman, N., 2010.

Extremotolerance in fungi: evolution on the edge. FEMS Microbiol. Ecol. 71,

2–11. doi:10.1111/j.1574-6941.2009.00794.x

Grant, W.D., 2004. Life at low water activity. Philos. Trans. R. Soc. London 359,

1249–1267. doi:10.1098/rstb.2004.1502

Gunde-Cimerman, N., Ramos, J., Plemenitas, A., 2009. Halotolerant and halophilic

fungi. Mycol. Res. 113, 1231–41. doi:10.1016/j.mycres.2009.09.002

Gunde-Cimerman, N., Sonjak, S., Zalar, P., Frisvad, J.C., Diderichsen, B., Plemenitaš,

A., 2003. Extremophilic fungi in arctic ice: a relationship between adaptation

to low temperature and water activity. Phys. Chem. Earth 28, 1273–1278.

doi:10.1016/j.pce.2003.08.056

Gunde-Cimerman, N., Zalar, P., 2014a. Extremely halotolerant and halophilic fungi

inhabit brine in solar salterns around the globe. Food Technol. Biotechnol. 52,

170–179.

22 Gunde-Cimerman, N., Zalar, P., 2014b. Extremely halotolerant and halophilic fungi

inhabit brine in solar salterns around the globe. Food Technol. Biotechnol. 52,

170–179.

Gunde-Cimerman, N., Zalar, P., de Hoog, S., Plemenitas, A., 2000. Hypersaline

waters in salterns — natural ecological niches for halophilic black yeasts.

FEMS Microbiol. Ecol. 32, 235–240.

Hammami, W., Fiori, S., Al Thani, R., Ali Kali, N., Balmas, V., Migheli, Q., Jaoua,

S., 2014. Fungal and aflatoxin contamination of marketed spices. Food Control

37, 177–181. doi:10.1016/j.foodcont.2013.09.027

Hocking, A.D., Pitt, J.I., 1980. Dichloran-glycerol medium for enumeration of

xerophilic fungi from low-moisture foods. Appl. Environ. Microbiol. 39, 488–

492.

Iacumin, L., Chiesa, L., Boscolo, D., Manzano, M., Cantoni, C., Orlic, S., Comi, G.,

2009. Moulds and ochratoxin A on surfaces of artisanal and industrial dry

sausages. Food Microbiol. 26, 65–70. doi:10.1016/j.fm.2008.07.006

Javor, B.J., 2002. Industrial microbiology of solar salt production. J. Ind. Microbiol.

Biotechnol. 28, 42–47. doi:10.1038/sj/jim/7000173

Jay, J.M., Loessner, M.J., Golden, D.A., 2008. Modern food microbiology, 7th ed.

Springer.

Lenth, R., Hervé, M., 2015. lsmeans: Least-Squares Means [WWW Document]. R

Package. version 2.16. URL http://cran.r-project.org/package=lsmeans

Mandeel, Q. a., 2005. Fungal contamination of some imported spices. Mycopathologia

159, 291–298. doi:10.1007/s11046-004-5496-z

23 O’Hara, R.B., 2009. How to make models add up—a primer on GLMMs. Ann. Zool.

Fennici 46, 124–137. doi:10.5735/086.046.0205

Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B.,

Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2015. Vegan:

community ecology package [WWW Document]. Packag. version 2.2-1. URL

http://cran.r-project.org/package=vegan

Papagianni, M., Ambrosiadis, I., Filiousis, G., 2007. Mould growth on traditional

greek sausages and penicillin production by Penicillium isolates. Meat Sci. 76,

653–7. doi:10.1016/j.meatsci.2007.01.018

Parks, C., 2014. Shaking up salt: Craft salt farmers go against the grain. Mod. Farmer.

Winter 2014/15. 11-13.

Pitt, J.I., Hocking, A.D., 2009. The ecology of fungal food spoilage, in: Fungi and

Food Spoilage. pp. 3–9. doi:10.1007/978-0-387-92207-2

Ponsone, M.L., Combina, M., Dalcero, A., Chulze, S., 2007. Ochratoxin A and

ochratoxigenic Aspergillus species in Argentinean wine grapes cultivated

under organic and non-organic systems. Int. J. Food Microbiol. 114, 131–135.

doi:10.1016/j.ijfoodmicro.2006.07.001

R Core Team, 2015. R: A language and environment for statistical computing. [WWW

Document]. URL http://www.r-project.org/.

Ramette, A., 2007. Multivariate analyses in microbial ecology. FEMS Microbiol.

Ecol. 62, 142–160. doi:10.1111/j.1574-6941.2007.00375.x

Rossman, A.Y., Samuels, G.J., Rogerson, C.T., Lowen, R., 1999. Genera of

Bionectriaceae, Hypocreaceae and Nectriaceae (Hypocreales, Ascomycetes).

24 Centraalbureau voor Schimmelcultures 42, 1–248.

Runa, F., Park, M.S., Pryor, B.M., 2009. Ulocladium systematics revisited: Phylogeny

and taxonomic status. Mycol. Prog. 8, 35–47. doi:10.1007/s11557-008-0576-y

Saitou, N., Nei, M., 1987. The neighbor-joining method: A new method for

reconstructing phylogenetic trees. Mol. Biol .Evol. 4, 406–425.

Samson, R.A., Mouchacca, J., 1975. Additional notes on species of Aspergillus,

Eurotium and Emericella from Egyptian desert soil. Antonie Van

Leeuwenhoek 41, 343–351.

Samson, R. A., Visagie, C.M., Houbraken, J., Hong, S.-B., Hubka, V., Klaassen,

C.H.W., Perrone, G., Seifert, K. A., Susca, a., Tanney, J.B., Varga, J.,

Kocsubé, S., Szigeti, G., Yaguchi, T., Frisvad, J.C., 2014. Phylogeny,

identification and nomenclature of the genus Aspergillus. Stud. Mycol. 78,

141–173. doi:10.1016/j.simyco.2014.07.004

Schoch, C.L., Seifert, K. A., Huhndorf, S., Robert, V., Spouge, J.L., Levesque, C.A.,

Chen, W., 2012. Nuclear ribosomal internal transcribed spacer (ITS) region as

a universal DNA barcode marker for Fungi. Proc. Natl. Acad. Sci. U. S. A.

109, 6241–6. doi:10.1073/pnas.1117018109

Schubert, K., Groenewald, J.Z., Braun, U., Dijksterhuis, J., Starink, M., Hill, C.F.,

Zalar, P., De Hoog, G.S., Crous, P.W., 2007. Biodiversity in the Cladosporium

herbarum complex (Davidiellaceae, Capnodiales), with standardisation of

methods for Cladosporium taxonomy and diagnostics. Stud. Mycol. 58, 105–

156. doi:10.3114/sim.2007.58.05

Sonjak, S., Ličen, M., Frisvad, J.C., Gunde-Cimerman, N., 2011a. Salting of dry-cured

25 meat — A potential cause of contamination with the ochratoxin A-producing

species Penicillium nordicum. Food Microbiol. 28, 1111–6.

doi:10.1016/j.fm.2011.02.007

Sonjak, S., Ličen, M., Frisvad, J.C., Gunde-Cimerman, N., 2011b. The mycobiota of

three dry-cured meat products from Slovenia. Food Microbiol. 28, 373–6.

doi:10.1016/j.fm.2010.09.007

Sørensen, L.M., Jacobsen, T., Nielsen, P.V., Frisvad, J.C., Koch, A.G., 2008.

Mycobiota in the processing areas of two different meat products. Int. J. Food

Microbiol. 124, 58–64. doi:10.1016/j.ijfoodmicro.2008.02.019

Steyn, P.S., 1969. The isolation, structure and absolute conFigureuration of secalonic

acid D, the toxic metabolite of Penicillium oxalicum. Tetrahedrom 326, 51–57.

Visagie, C.M., Hirooka, Y., Tanney, J.B., Whitfield, E., Mwange, K., Meijer, M.,

Amend, a. S., Seifert, K. A., Samson, R. A., 2014a. Aspergillus, Penicillium

and Talaromyces isolated from house dust samples collected around the world.

Stud. Mycol. 78, 63–139. doi:10.1016/j.simyco.2014.07.002

Visagie, C.M., Houbraken, J., Frisvad, J.C., Hong, S.-B., Klaassen, C.H.W., Perrone,

G., Seifert, K. A., Varga, J., Yaguchi, T., Samson, R. A., 2014b. Identification

and nomenclature of the genus Penicillium. Stud. Mycol. 78, 343–371.

doi:10.1016/j.simyco.2014.09.001

Vytrasová, J., Pribánová, P., Marvanová, L., 2002. Occurrence of xerophilic fungi in

bakery gingerbread production. Int. J. Food Microbiol. 72, 91–6.

White, T.J., Bruns, S., Lee, S., Taylor, J., 1990. Amplification and direct sequencing

of fungal ribosomal RNA genes for phylogenetics, in: PCR Protocols: A Guide

26 to Methods and Applications. pp. 315–322.

Wolfe, B.E., Button, J.E., Santarelli, M., Dutton, R.J., 2014. Cheese rind communities

provide tractable systems for in situ and in vitro studies of microbial diversity.

Cell 158, 422–433. doi:10.1016/j.cell.2014.05.041

Alimentarius, C., 1985. Codex standard for food grade salt [WWW Document]. Codex

stan 150, 1985. URL http://www.fao.org/fao-who-codexalimentarius/sh-

proxy/ar/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites

%252Fcodex%252FStandards%252FCODEX%2BSTAN%2B150-

1985%252FCXS_150e.pdf (accessed 6.20.17).

Zajc, J., Dzeroski, S., Kocev, D., Oren, A., Sonjak, S., Tkavc, R., Gunde-Cimerman,

N., 2014. Chaophilic or chaotolerant fungi: a new category of extremophiles?

Front. Microbiol. 5, 1–15. doi:10.3389/fmicb.2014.00708

Zalar, P., de Hoog, G.S., Schroers, H.-J., Crous, P.W., Groenewald, J.Z., Gunde-

Cimerman, N., 2007. Phylogeny and ecology of the ubiquitous saprobe

Cladosporium sphaerospermum, with descriptions of seven new species from

hypersaline environments. Stud. Mycol. 58, 157–83.

doi:10.3114/sim.2007.58.06

CHAPTER 2

PAECILOMYCES ROT: A NEW APPLE DISEASE2

INTRODUCTION

Pecilomyces niveus Stolk & Samson (Byssochlamys nivea Westling) (Ascomycota,

Eurotiomycetes, Eurotiales, Thermoascaceae) is an important food spoilage fungus

(Houbraken and Samson 2017). It causes disintegration of thermally processed foods while on the shelf, and can produce the federally-regulated mycotoxin patulin

(Houbraken et al. 2006; Puel et al. 2010; Sant’Ana et al. 2008). Not only does it produce conidia and chlamydospores, but being homothallic, it readily produces abundant, durable ascospores that can survive high heat (Samson et al. 2009).

Ascospores of P. niveus can survive temperatures of 90°C, leading to spoilage of thermally processed syrups and canned fruits (Beuchat and Toledo 1977; Tournas

1994). This hard-to-kill species and its relatives in the genus Paecilomyces grow in harsh conditions, including on leather, in alcohol, and under low oxygen conditions

(Beuchat and Rice 1979; Beuchat and Toledo 1977; Brown and Smith 1957; Taniwaki et al. 2009, 2010).

Paecilomyces niveus can grow in apple products containing up to 40 g 100 ml-1 added sucrose (40% weight by volume) and at a water activity of 0.92 (Beuchat and

2 Biango-Daniels and Hodge. 2018. Paecilomyces rot: A new apple disease. Accepted in J Plant Disease. https://doi.org/10.1094/PDIS-12-17-1896-RE

Toledo 1977). A water activity this low inhibits the growth of most other spoilage organisms by limiting the water availability and creating an osmotic gradient that dehydrates growing fungi (Beuchat and Toledo 1977; Grant 2004). Spoilage of thermally processed foods by P. niveus has been speculated to result from soil contamination (Brown and Smith 1957; King et al. 1969). The fungus has been found in orchard soils and on decaying fruit and vegetable matter, but its ecology is not well understood (Beuchat and Rice 1979; Splittstoesser et al. 1971).

One clue to the ecology of P. niveus may be its production of patulin, which is not only a mycotoxin, but also a well-established virulence factor. It enhances the ability of the distantly related fungus Penicillium expansum to invade apples (Sanzani et al. 2012; Snini et al. 2016), leading to Penicillium blue mold rot, an important post- harvest disease of apples. Some studies have raised questions about the potential of P. niveus to act as an apple pathogen, which could give it a direct route into apple-based foods (Dombrink-Kurtzman and Engberg 2006; Salomão et al. 2014; Splittstoesser et al. 1971).

This study aimed to elucidate the ecology of P. niveus. A selective isolation protocol was used to survey for it in orchard soils and culled apples, and the ability of

P. niveus to act as a post-harvest pathogen of apples was tested.

MATERIALS AND METHODS

Identification methods.

Putative isolates of P. niveus were morphologically identified using the

Ascomycete key developed by Samson et al. (2000). To distinguish P. niveus from closely related species, genomic DNA was extracted and the internal transcribed spacer (ITS) region of the ribosomal DNA, or the BenA region of the β-tubulin gene was amplified and sequenced. Mycelium was harvested from P. niveus cultures grown on potato dextrose agar (PDA: Criterion). Genomic DNA was extracted using the

PowerSoil DNA Isolation Kit (MoBio), following the recommended protocol, and stored in elution buffer at -20°C. Genomic DNA was the target for PCR amplification of the ITS region of the rDNA repeat, a widely accepted barcode for fungi, using forward ITS5 (5’-GGAGTAAAAGTCGTAACAAGG -3’) and reverse ITS4 (5’-

TCCTCCGCTTATTGATATGC-3’) primers (White et al. 1990; Schoch et al. 2012).

The BenA region of the β-tubulin gene was amplified using forward Bt2a (5’-

GGTAACCAAATCGGTGCTGCTTTC -3’) and reverse Bt2b (5’-

ACCCTCAGTGTAGTGACCCTTGGC-3’) primers (Visagie et al. 2014).

Sigma-Aldrich REDTaq genomic DNA polymerase with MgCl2 was used in a

DNA Engine PTC-200 thermal cycler (MJ Research). The ITS amplification was programmed as follows: initial denaturation 95°C/2 min, then 26 cycles of 94°C/1 min, 55°C/1 min, and 72°C/3 min, and final extension of 72°C/15 min. The BenA region was amplified using the program: initial denaturation 95°C/2 min, then 30 cycles of 94°C/1 min, 57°C/1 min, and 72°C/3 min, and final extension of 72°C/15

min. PCR amplicons were cleaned with the E.N.Z.A. Cycle-Pure kit (Omega Bio-tek), following manufacturer instructions. Amplicons were sequenced with the forward

PCR primers, using an Applied Biosystems Automated 3730xl DNA analyzer, at the

Cornell University Biotechnology Resource Center. The sequences were compared with reference sequences in NCBI using BLASTn (Agostino, 2012; Altschul et al.,

1990), and they were subject to phylogenetic analyses.

Phylogenetic analyses.

To confirm morphological identification of isolates and place them phylogenetically, two phylogenetic analyses were conducted: one using the partial - tubulin alone; the other using a concatenation of the partial -tubulin and ITS regions for isolate CO7 (this study). The identities of isolates used in this study were inferred by comparing their phylogenetic relationship to those of Samson et al. (2009).

Accession numbers provided in Houbraken et al. (2008) and Samson et al. (2009) were used to acquire sequences for the partial -tubulin and ITS regions from NCBI

GenBank (Table 2.1). Only isolate sequences that were used in partial -tubulin and

ITS region tree construction of Samson et al. (2009) were used in the phylogenetic analysis (Table 2.1), apart from Paecilomyces fulvus (Byssochlamys fulva) CBS

113954, for which only the partial -tubulin sequence was included.

For the partial -tubulin analysis, sequences from putative P. niveus isolates generated in this study were aligned to those used in Samson et al. (2009) using

MUSCLE (Edgar 2004) via the EMBL-EBI on-line bioinformatics resource (Li et al.

2015), with poorly aligned, divergent regions trimmed from the alignment using

GBlocks version 0.91b (Castresana 2000), specifying the less stringent criteria. The total number of taxa and characters in the partial -tubulin analysis were 47 and 437 respectively. For the concatenated partial -tubulin and ITS region analysis, the partial

-tubulin and ITS regions of isolate CO7 (this study) were aligned to those isolates used by Samson et al. (2009), for which both gene sequences were available. Each alignment was subsequently trimmed as previously described. The partial -tubulin and ITS region alignments were concatenated using SequenceMatrix v1.8 (Vaidya et al. 2011). The total number of taxa and characters in the concatenated partial β- tubulin and ITS region analysis were 31 and 984 respectively.

For both the partial -tubulin and the concatenated partial -tubulin and ITS region analyses, trees were constructed by Bayesian inference using MrBayes v3.2.6

(Ronquist et al. 2012). The partial β-tubulin analysis was conducted using K80+G model, selected by jModelTest 2.1 (Darriba et al. 2012) with the Bayesian information criterion. The Bayesian model averaging approach of rjMCMC was applied to the concatenated partial β-tubulin and ITS region analyses (Huelsenbeck et al. 2004). In the concatenated analysis, each gene was considered a separate partition, allowing for independent estimation of substitution rates, character state frequencies, gamma shape distributions, and proportions of invariable sites per partition. Posterior probabilities at branch nodes were inferred using two runs of 1,000,000 generations, with 25% burn- in. Phylogenetic trees were visualized in R v3.4.2, using the Ape package (Paradis et al. 2004), in the RStudio v1.1.383 environment, rooted with Thermoascus crustaceus as the outgroup.

Table 2.1. Table if isolates and accession numbers used in phylogenetic analyses. *All CBS isolate identifications are per Table 1 in Samson et al. (2009). Species identifications for isolates derived in the present study are based on phylogenetic affiliation with those of Paecilomyces niveus (Byssochlamys nivea) sensu Samson et al. (2009).

Species Isolate Source ITS β-tubulin identification* Byssochlamys Samson et al. CBS 113954 - FJ389985 fulva 2009 Byssochlamys Samson et al. CBS 132.33 FJ389939 FJ389988 fulva 2009 Byssochlamys Samson et al. CBS 135.62 FJ389943 FJ389989 fulva 2009 Byssochlamys Samson et al. CBS 146.48 FJ389940 FJ389986 fulva 2009 Byssochlamys Samson et al. CBS 604.71 FJ389941 FJ389997 fulva 2009 Byssochlamys Samson et al. CBS 110378 FJ389946 FJ390006 lagunculariae 2009 Byssochlamys Samson et al. CBS 373.70 FJ389944 FJ389995 lagunculariae 2009 Byssochlamys Samson et al. CBS 696.95 FJ389945 FJ389996 lagunculariae 2009 Byssochlamys Samson et al. CBS 100.11 FJ389934 FJ389999 nivea 2009 Byssochlamys Samson et al. CBS 113245 FJ389936 FJ389998 nivea 2009 Byssochlamys Samson et al. CBS 133.37 FJ389935 FJ390000 nivea 2009 Paecilomyces CO7 This study MG008851 MG008852 niveus

Paecilomyces 100-10 This study - MG008853 niveus Paecilomyces 101-3 This study - MG008854 niveus Paecilomyces 102-1 This study - MG008855 niveus Paecilomyces 104-22 This study - MG008856 niveus Paecilomyces 106-3 This study - MG008857 niveus Paecilomyces 107-1 This study - MG008858 niveus Paecilomyces 108-11 This study - MG008859 niveus

Table 2.1 Continued Paecilomyces 109-21 This study - MG008860 niveus Paecilomyces 110-3 This study - MG008861 niveus Paecilomyces 111-1 This study - MG008862 niveus Paecilomyces 112-22 This study - MG008863 niveus Paecilomyces 125-31 This study - MG008864 niveus Paecilomyces 140-11 This study - MG008865 niveus Paecilomyces 141-11 This study - MG008866 niveus Paecilomyces 146-311 This study - MG008867 niveus Byssochlamys Houbraken et CBS 101075 EU037051 EU037069 spectabilis al. 2008 Byssochlamys Houbraken et CBS 102.74 EU037055 EU037073 spectabilis al. 2008 Byssochlamys Houbraken et CBS 121581 EU037062 EU037080 spectabilis al. 2008 Byssochlamys Samson et al. CBS 338.51 FJ389930 FJ390007 spectabilis 2009 Byssochlamys Samson et al. CBS 374.70 FJ389933 FJ390008 zollerniae 2009 Paecilomyces Houbraken et CBS 370.70 EU037050 EU037068 brunneolus al. 2008 Paecilomyces Samson et al. CBS 110429 FJ389932 FJ389991 divaricatus 2009 Paecilomyces Samson et al. CBS 284.48 FJ389931 FJ389992 divaricatus 2009 Paecilomyces Samson et al. CBS 113247 FJ389921 FJ390009 formosus 2009 Paecilomyces Samson et al. CBS 296.93 FJ389928 FJ389994 formosus 2009 Paecilomyces Samson et al. CBS 371.70 FJ389920 FJ389982 formosus 2009

Table 2.1 Continued Paecilomyces Samson et al. CBS 372.70 FJ389926 FJ389990 formosus 2009 Paecilomyces Samson et al. CBS 628.66 FJ389927 FJ389983 formosus 2009 Paecilomyces Samson et al. CBS 990.73B FJ389929 FJ389993 formosus 2009 Paecilomyces Samson et al. CBS 251.55 FJ389951 FJ390002 saturatus 2009 Paecilomyces Samson et al. CBS 323.34 FJ389947 FJ390005 saturatus 2009 Paecilomyces Samson et al. CBS 368.70 FJ389948 FJ390001 saturatus 2009 Paecilomyces Samson et al. CBS 492.84 FJ389949 FJ390003 saturatus 2009 Paecilomyces Samson et al. CBS 990.73A FJ389950 FJ390004 saturatus 2009 Thermoascus Samson et al. CBS 181.67 FJ389925 FJ389981 crustaceus 2009

Detection of Paecilomyces niveus in New York orchard soil.

To determine whether P. niveus was present in New York state orchard soils, in the summer of 2016, soil samples were taken from the top six inches of soil in orchards across the state by Cornell Cooperative Extension agents and independent consultants. This depth was chosen because it represents the soil that may contact apples during harvest or if they are on the ground. Soils were sampled and mailed to

Cornell University, refrigerated (4°C) upon receipt, and processed within 2 weeks. In total 32 different orchards were sampled across New York state. Due to miscommunication, 4 of the orchards were sampled in multiple places, which lead to

48 soil samples being received by the Hodge Lab.

Soils were thermally processed by mixing them with sterile deionized (DI) water at 1:2 weight:volume. The resulting dilute soil samples were placed in 2 ml microcentrifuge tubes and heated in a water bath at 70°C for 30 minutes. This process reduces the number of microorganisms present and has been shown to increase germination of P. niveus ascospores (Splittstoesser et al. 1970). For each cooled sample, 200 µl was spread with a sterile L-shaped rod on five PDA Petri plates (90 mm) supplemented with 30 µg ml -1 streptomycin sulfate. After a 3-week incubation at

25°C all emerging colonies with a morphology consistent with P. niveus were subcultured on PDA at 25°C (Beuchat and Pitt 2001). As described above (See

Identification and Phylogenetic analyses methods), isolates were morphologically identified and the identity was confirmed by phylogenetic analysis of sequences of the

BenA region of the partial β-tubulin gene, a recognized fungal barcode that differentiates closely related species of Paecilomyces (Schoch et al. 2012).

Isolation of Paecilomyces niveus from culled apples.

In addition to surveying for P. niveus in orchard soils, stored apples with signs of post-harvest diseases were sampled to detect any P. niveus present. In the spring of

2017, Cornell Orchards staff selected apples displaying symptoms of post-harvest diseases as bins were pulled from cold storage. Cornell Orchards, located in Ithaca,

New York, is a teaching and retail facility that grows apples for fresh market and cider production. The apples, which were ‘Mutsu’, ‘Goldrush’, ‘Empire’ and ‘Shizuka,’ had been stored at 1°C since harvest the previous autumn.

Each week, for 5 weeks, a 5-gallon bucket of culled apples was received and apple lesions were dissected to harvest infected tissue. To reduce the possibility of surface soil contaminants, apples were first surface sanitized with 70% ethanol. The pooled weekly sample, weighing between 40-51g, was frozen and stored at -20°C for

2-3 months. Additionally, a separate sample from an apple, inoculated with P. niveus in the lab as described below (see Inoculation method), was aseptically dissected and its infected tissue was included as a positive control. Its tissue was subjected to approximately the same freezing time.

To isolate any P. niveus present, tissue samples taken from culled apples were combined with sterile DI H2O 1:1 weight:volume. The suspensions were homogenized in sterile stomacher bags (Seward Stomacher 400 circulator) for 2 minutes at 230

RPM (Beuchat and Pitt 2001). Next, samples were divided among five 10 ml glass test tubes; four were incubated at 75°C for 30 minutes in a closed water bath, while the fifth was not heated. The unheated sample was used to estimate the background number of non-heat resistant fungi present. Duplicate aliquots (100 µl) of apple

homogenate from each of the four heated test tubes per sample were spread on Petri dishes (90 mm) of Rose Bengal agar (HIMEDIA). Dilutions from 10-1 to 10-4 were made of the unheated samples, and aliquots (100 µl) of each dilution were spread on

Rose Bengal agar. All Petri plates were incubated at 30°C for one month (dark), and emerging colonies were counted.

Koch’s postulates.

To test whether P. niveus can cause disease in apples, the four steps of Koch’s postulates were completed: (i) Paecilomyces niveus isolate CO7 was isolated from a rotting apple discovered at Cornell Orchards in the fall of 2014; (ii) It was grown in pure culture and morphologically and genetically identified (Samson et al. 2000); (iii)

The culture of CO7 was used to inoculate healthy apples through a wound (see

Inoculation method) and disease symptoms were observed on the inoculated apples using a dissecting microscope (Olympus SZX12); (iv) The fungus was reisolated from within the lesions of infected apples and confirmed to be P. niveus, using morphological and genetic approaches.

Inoculation method.

Fifteen ‘Gala’ and fifteen ‘Golden Delicious’ apples were sanitized (30 s in

70% ethanol, 2 min in 1% sodium hypochlorite, 15 s in 70% ethanol), and dried in a biological safety cabinet (Geisler 2013; van der Walt et al. 2010). Following sanitization, apples were numbered and inoculated with P. niveus CO7, which was identified morphologically (Samson et al. 2000), and confirmed through phylogenetic

analysis of sequences of the ITS and BenA regions (MG008851, MG008852) (See

Phylogenetic analyses).

Colonized toothpicks were used to wound and inoculate mature ‘Gala’ and

‘Golden Delicious’ apples. These apple varieties were selected because they are among the top ten commercial cultivars (US Apple Association). Toothpicks were inoculated with P. niveus CO7 as described for Penicillium species on apples (van der

Walt et al. 2010). Briefly, toothpick pieces (16 mm) were sterilized by autoclaving them five times in DI water, which was replaced between cycles. They were next transferred to potato dextrose broth (Teknova) and autoclaved a final time. Toothpick pieces were aseptically transferred to PDA Petri dishes (90 mm), containing agar plugs

(5 mm) of a 1-week-old P. niveus CO7 culture. After 2 weeks of incubation (30°C, dark), toothpick pieces were well colonized by P. niveus. Control toothpicks were similarly treated except they were placed on uncolonized PDA Petri dishes (90 mm).

Toothpick pieces were inserted at four points, each halfway between the stem and calyx, to wound the sanitized apples. Two treatments were applied to each apple: 1) two opposite wounds inoculated with P. niveus CO7; and 2) two opposite wounds that received control toothpicks. Inoculated ‘Gala’ and ‘Golden Delicious’ apples were incubated in separate dark moist chambers (25°C, ≥ 95% humidity) for 3 weeks.

Disease characterization.

To describe the disease symptoms that developed following inoculation with P. niveus, ‘Gala’ and ‘Golden Delicious’ apples in moist chambers were monitored.

Lesions were measured three times a week during the 3-week incubation period. Using

digital calipers (VWR Carbon Fiber Composite), lesion diameters were measured at two perpendicular angles, and averaged. To allow for comparisons of each wound over time, individual wounds and apples were tracked with an identification number.

To eliminate the possibility of lesion size being influenced by contaminating fungi, any apples that showed signs of contamination during incubation, specifically by

Penicillium species, were removed from the experiment. Several apples of each variety were maintained in moist chambers for an extended period to observe disease symptoms at an advanced stage.

Statistical analysis.

Two separate linear mixed effects models, one at day 7 and one at day 22, were used to test for significant differences between lesion diameters for both treatments

(control or inoculated) on both varieties of apples. Fixed effects of apple variety

(which was confounded with chamber), treatment, and their interaction were included and a random effect of apple identification number was included, which accounted for the multiple wounds on each apple. These models were executed using the lme4 and lmerTest packages in R statistical programming software. Post-hoc comparisons were carried out using the lsmeans package. The data were visualized using the ggplot2 package.

RESULTS

Identification and phylogeny.

BLASTing (Blastn) of the sequenced regions of CO7 (MG008851,

MG008852) showed that CO7 closely matched P. niveus isolates in NCBI GenBank.

The ITS and BenA regions of CO7 were concatenated and the placement of the isolate within the P. niveus clade further confirmed its identity (Figure 2.1). The P. niveus isolates from New York state orchard soils were morphologically identified, and their identities were supported by phylogenetic analysis (Figure 2.2). All isolates from soil clustered with P. niveus (B. nivea) that were predominated isolated from spoiled foods

(Samson et al. 2009) (Figure 2.2).

Paecilomyces niveus in New York orchard soils and culled apples.

Fifteen Paecilomyces niveus were isolated from 11 of the 32 New York orchards sampled in 2016 (Table 2.2). Due to unintentional sampling within several orchards, Paecilomyces niveus was isolated at multiple places in the orchards 5 (n=2) and 6 (n=4). This provides evidence that P. niveus may be widespread in the soil throughout orchards (Table 2.2).

Figure 2.1. Consensus topology and branch lengths of concatenated partial β-tubulin and ITS sequences of 31 taxa, including Paecilomyces niveus (Byssochlamys nivea) isolate CO7 (this study, in bold), rooted with Thermoascus crustaceus as an outgroup. Tree constructed with MrBayes v3.2.6 using model averaging (rjMCMC). Support values at nodes represent posterior probabilities, scale bar shows the number of substitutions per site.

Figure 2.2. Consensus topology and branch lengths of partial β-tubulin sequences of 47 taxa, including Paecilomyces niveus (Byssochlamys nivea) isolates from New York orchards, rooted with Thermoascus crustaceus as an outgroup. Species designations are described by Samson et al. 2009 (Table 2.1). Tree constructed using MrBayes v3.2.6 with K80+g model. Support values at nodes represent posterior probabilities, scale bar shows the number of substitutions per site. Isolates from this study are in bold.

After one-month incubation at 30°C, three samples from the not-heat-treated culled apples had fungal colonies that were counted. Fungal counts for these samples ranged from 1,000-22,000 CFU ml-1. Penicillium species were most common on these plates. The other two not-heat-treated culled apple samples showed no fungal growth.

Among the five heat-treated samples, only one isolate of a heat resistant mold was found. It was determined not to be a Paecilomyces species. As expected, Paecilomyces niveus was present in the positive control sample, but it was not detected after freezing and heat treatment of that sample.

Table 2.2. Origins of P. niveus isolates used in this study from soil in New York orchards. Location Location Isolate Orchard Isolate Orchard County Town County Town 100-10 1 Tompkins Ithaca 110-3 6 Clinton Peru 101-3 2 Clinton Geneva 111-1 6 Clinton Peru 102-1 3 Clinton Geneva 112-22 7 Clinton Chazy 104-22 4 Clinton Peru 125-31 8 Wayne Sodus 106-3 5 Clinton Peru 140-11 9 Dutchess Shenandoah 107-1 5 Clinton Peru 141-11 10 Ulster New Paltz 146- 108-11 6 Clinton Peru 11 Wayne Newark 311 109-21 6 Clinton Peru

Paecilomyces niveus causes disease in apples.

Isolate CO7, taken from rotting apples at Cornell Orchards in 2014, was morphologically identified as P. niveus and its identity is supported by phylogenetic analysis (Figure 2.1). It was grown in pure culture and then used to inoculate healthy apples. After the infected apples began to show symptoms of disease, a fungus was reisolated, grown in pure culture and identified as P. niveus. Koch’s postulates were completed by using the fungus to infect healthy apples.

Characterization of Paecilomyces rot.

Apples inoculated with P. niveus developed circular lesions that radially expanded from the point of inoculation (Illustration 2.1 A-B). Initially, lesions were tan to brown, resembling a bruise. As lesions expanded, they became brown with a lighter brown center. Within 12 days post-inoculation (dpi), circular lesions enlarged and displayed a series of faint concentric rings, alternating from tan to brown

(Illustration 2.1 A), which seemed more visible on ‘Golden Delicious’ than on ‘Gala’ apples. Within the infected fruit medium brown colored sections of necrotic tissue progressed through the mesocarp toward the endocarp in a tapered cone-shape, slightly bowed outward into the surrounding mesocarp (Illustration 2.1 C). This discoloration was firm and spongy, and other than the color, was not easily separable from the healthy tissue (Illustration 2.1 C).

Illustration 2.1. Disease symptoms caused by Paecilomyces niveus (Byssochlamys nivea) after incubation in dark moist chambers (25°C, ≥ 95% humidity). (A-B) External symptoms 12 dpi on (A) ‘Golden Delicious,’ and (B) ‘Gala’ fruit. Insets showing a control apple of each variety. Scale bars are 1.8 mm. (C) Cross section of ‘Gala’ apple, 12 dpi, showing representative cone-shaped internal rot. (D) Longitudinal section of ‘Gala’ apple, showing internal marbling at 44 dpi. (E-F) ‘Gala’ apple with advanced Paecilomyces Rot at 44 dpi. (F) Close-up view of tufts of white mycelia produced on the surface of an infected apple.

After an extended incubation period (44 days), advanced disease symptoms included the complete discoloration of the apple epidermis to light tan-brown

(Illustration 2.1 E-F). The surface of these apples was observed to be sparsely covered by branching white hyphae, concentrated around lenticels in floccose tufts (Illustration

2.1 F). After 44 days, the internal necrotic tissue extended to the cores of both apple varieties (Illustration 2.1 D), which were filled with white mycelium and asci containing ascospores (Illustration 2.2). At this late stage of necrosis, tissue was marbled in shades of tan and dark brown (Illustration 2.1 D).

Illustration 2.2. Representative image of Paecilomyces niveus (Byssochlamys nivea) asci containing ascospores, observed in the cores of infected apples. The scale bar is 10µm.

Size of Paecilomyces niveus lesions.

Wounded apples that were inoculated had lesions that grew rapidly. The average diameters of treatment lesions, measured 7 days after inoculation, were 13.3 ±

0.4 mm standard error (n=30) and 17.1 ± 0.3 mm standard error (n=30) on the ‘Gala’ and the ‘Golden Delicious’ respectively (Figure 2.3). The average diameters of treatment lesions, measured 22 days after inoculation, were 47.6 ± 0.7 mm standard error (n=24) and 43.2 ± 0.9 mm standard error (n=26) on the ‘Gala’ and the ‘Golden

Delicious’ respectively (Figure 2.3). The diameter of control lesions remained near the size of the original inoculation wound (1.8 mm) throughout the study. The average diameter of control lesions, measured 7 days after inoculated, were 2.0 ± 0.3 mm standard error (n=30) and 2.4 ± 0.4 mm standard error (n=30) on the ‘Gala’ and

‘Golden Delicious’ respectively. After the 22-day incubation, control lesions increased from the initial 1.8 mm diameter of the toothpick to an average of 3.4 ± 0.9 mm standard error (n=24) and 3.2 ± 0.9 mm standard error (n=26) on the ‘Gala’ and the

‘Golden Delicious’ respectively (Figure 2.3).

The diameter of Paecilomyces rot treatment lesions, on both the ‘Gala’ and

‘Golden Delicious’ were significantly larger than control lesions on days 7 and 22

(P<0.0001). Although confounded by chamber, treatment lesions on the ‘Gala’ were smaller than those on ‘Golden Delicious’ early in the disease (P<0.0001, day 7), however, this trend reversed by day 22, when ‘Gala’ treatment lesions were larger than those on the ‘Golden Delicious’ (P=0.005) (Figure 2.3). There were no significant differences in control lesion diameters on days 7 or 22 on ‘Gala’ or ‘Golden

Delicious.’ Over the course of the experiment, six ‘Gala’ and four ‘Golden Delicious’

apples were removed from their respective incubation chambers because they showed signs of being colonized by other fungi. Once removed, these apples were excluded from further measurements and statistical analysis.

Figure 2.3. Increase, over a 22-day incubation (25°C, ≥ 95% humidity), in the mean diameter of lesions (± 1 standard error) on ‘Gala’ and ‘Golden Delicious’ apples inoculated with Paecilomyces niveus (Byssochlamys nivea). Treatment lesions were significantly larger than control lesions on the ‘Gala’ and ‘Golden Delicious’ at days 7 and 22 (P<0.0001).

DISCUSSION

This is the first report that P. niveus can infect apples, which are affected by many other post-harvest fungal diseases (Houbraken et al. 2006; Jurick and Cox 2016;

Sutton et al. 2014). At least 10–15% of apples are lost during storage because of disease; more in developing markets (Jurick and Cox 2016). In the U.S. post-harvest

losses cost an estimated $6.5 million annually (Jurick and Cox 2016). This work demonstrates that P. niveus is commonly found in orchard soils in New York, and can cause a novel post-harvest fruit disease, Paecilomyces rot, on two popular varieties of apples (Piecková et al. 1994; Splittstoesser et al. 1971).

Prior research suggested that soil was the source of food contamination, and indeed P. niveus was present in many New York orchard soils (Table 2.2). Not only was P. niveus in more than a third of the orchards sampled, but isolates were found at multiple points in two of those orchards (Table 2.2). This finding demonstrates that P. niveus is widespread across New York and that it may be common throughout orchards. All P. niveus isolates that were found in orchard soils shared a clade with other isolates of this species in the phylogenetic analysis (Figure 2.1). Very little is known about the ecology of P. niveus and its relatives, however none of these food spoilage species have been previously described as plant pathogens. Paecilomyces species are known for the food spoilage problems they cause in a variety of thermally processed foods, but their biology and niches in nature are virtually unstudied.

Paecilomyces niveus and some P. saturatus are the only species in the genus known to produce patulin (Samson et al. 2009). Perhaps this toxin is responsible for the ability of P. niveus to infect apples, as patulin is a virulence factor in apples in the distantly related blue rot pathogen, Penicillium expansum (Ascomycota, Eurotiomycetes,

Eurotiales, Trichocomaceae) (i.e. they are in different families) (Baert et al. 2007;

Sanzani et al. 2012).

Examining the P. niveus isolates found in this study in New York orchard soil and on a decaying apple (CO7) (Table 2.1) within a phylogenetic context, the P.

niveus that were isolated were in the same clade as P. niveus cultured from food environments (Figure 2.1), including pasteurized fruit juice in Europe (CBS 113245) and milk in North America (CBS 133.37) (Samson et al. 2009). Paecilomyces fulvus spoilage isolates, cultured from bottled fruit and fruit juice (CBS 132.33, CBS

135.62), were sister to P. niveus from orchard soil in our analysis, in agreement with previous work (Figure 2.1) (Samson et al. 2009). Because P. niveus and P. fulvus are close relatives that similarly spoil thermally processed fruits, future studies should evaluate whether P. fulvus can infect the fruits that it often spoils.

We hypothesize that infected fruit may be an overlooked source of spoilage of heat-treated apple products, when P. niveus survives thermal processing. The harmful practice of using “drops” (fallen apples) must be discouraged. Not only could “drops” passively transport spores in adhering soil particles, but apples may also be wounded and infected through soil contact. The ascospores that develop in infected fruit (Figure

3), may be able to survive thermal processing. Future work should ask whether fruits with Paecilomyces rot introduce this spoilage organism to thermally processed apple products.

In addition to spoiling food, P. niveus produces the mycotoxin patulin in food

(Moake et al. 2005; Sant’Ana et al. 2008). Patulin is not eliminated by thermal processing and is a major concern for the apple industry. Patulin acute toxicity includes convulsions, sub-acute toxicity can include gastrointestinal disorders, and there are concerns that it is carcinogenic (Puel et al. 2010). For these reasons, patulin is limited to 50 µg L-1 by the US Food and Drug Administration; to 50 ppb by Health

Canada; and to 10–25 µg L-1 by the European Union in apple juices and other

products, of which children are the biggest consumers and most at risk from exposure

(Bennett and Klich 2003; Health Canada 2018; Moake et al. 2005; US FDA 2001).

Although variety differences between ‘Gala’ and ‘Golden Delicious’ were confounded by chamber effects in our study, hints of differences in lesion size and symptoms between the varieties could have implications for contamination. Blue rot lesion diameter positively correlates with patulin accumulation in apples (Sant’Ana et al.

2008). Future research should investigate the accumulation of patulin in apples during the development of Paecilomyces rot and further evaluate disease progression of different apple varieties.

Paecilomyces rot has characteristics that make it difficult to distinguish from other apple diseases: bull’s-eye rot, apple ring rot, black rot, and bitter rot. It may have been overlooked because it is masquerading as these other diseases. It causes lesions like those seen in early stages of black rot, caused by Diplodia seriata, and bull’s-eye rot or apple ring rot, caused by Neofabraea species and Botryosphaeria dothidea respectively (Sutton et al. 2014). The internal symptoms of Paecilomyces rot resemble bitter rot disease, caused by Colletotrichum gloeosporioides, however in cross-section the infected cone-shaped tissue has edges that are more outwardly bowed, U-shape

(Phillips et al. 2007; Sutton et al. 2014) (Illustration 2.1 C). Paecilomyces rot symptoms, especially on ‘Golden Delicious’ apples, were reminiscent of bull’s-eye rot or black rot symptoms on apple fruit (Sutton et al. 2014). Like both rots, the lesions caused by P. niveus were slightly flattened and firm (Illustration 2.1 A-B). Unlike bull’s-eye and black rots, P. niveus produces no external fruiting bodies, such as pycnidia, pseudothecia, or acervuli (Illustration 2.1 F). Although possibly confounded

by a chamber effect in our experiment, surface hyphae were more often observed on

‘Gala’ than ‘Golden Delicious’ at the site of inoculation; the lenticels on ‘Gala’ were often colonized with floccose, white mycelia with asci (Illustration 2.1 B). The sparse white surface mycelium noted within Paecilomyces rot lesions (Illustration 2.1 B), distinguished it from bitter rot, which is known for copious quantities of conidia produced on the surface of lesions (Sutton et al. 2014). Compared to post-harvest diseases such as blue mold, which has a characteristic musty odor (Sutton et al. 2014), apples infected with P. niveus smelled sweet.

Another reason that P. niveus has not been previously described as an apple pathogen or isolated from apple storage environments could be due to sampling methods that are biased for fast growing Penicillium species (Amiri and Bompeix

2005). To test this idea, we attempted to isolate P. niveus from culled apples. After a recommended heat treatment to reduce other species and activate P. niveus ascospores

(Houbraken and Samson 2006), no P. niveus was recovered, even from apple tissue known to be infected. This finding may indicate that the temperature used was too high, or that combined freezing and heating damaged P. niveus. Younger ascospores are known to be less resilient (Bayne and Michener 1979; Dijksterhuis 2007). If P. niveus was present in the culled apples at low concentrations, we may have failed to isolate it because of low sample weights. New post-harvest sampling approaches, involving the addition of a mild heat treatment to increase germination, might be more successful (Yates et al. 1968).

Paecilomyces niveus is both important and unusual, as it infects apples, produces patulin, causes spoilage, and resists high temperatures, creating multiple

risks for processed apple products. To avert food safety problems caused by P. niveus, apple products may benefit from higher processing temperatures. More research is needed on the incidence of Paecilomyces rot, to trace a pathway from infected fruit to spoiled processed foods, and to investigate patulin production in infected apples.

ACKNOWLEDGEMENTS

We are grateful to Cornell Cooperative Extension agents, apple consultants, and orchard owners for their help in soil sampling. We thank the Cornell University

Orchards for their support of this work; Erika Mudrak of the Cornell Statistical

Consulting Unit for her assistance with data analyses; Jonathan Gonzalez for his assistance with phylogenetic analyses; and Kerik Cox for his advice on comparable diseases. This work was supported by the USDA National Institute of Food and

Agriculture, Hatch project 1002546, and by a Cornell CALS Arthur Boller Apple

Research Grant. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the National Institute of Food and Agriculture (NIFA) or the United States Department of Agriculture (USDA).

REFERENCES

Agostino, M., 2012. Introduction to the BLAST Suite and BLASTN. In: Practical

Bioinformatics, Garland Science. pp.47-71.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. 1990. Basic local

alignment search tool. Journal of molecular biology. 215(3):403-410.

Amiri, A., and Bompeix, G. 2005. Diversity and population dynamics of Penicillium

spp. on apples in pre- and postharvest environments: Consequences for decay

development. Plant Pathol. 54:74–81.

Baert, K., Devlieghere, F., Flyps, H., Oosterlinck, M., Ahmed, M.M., Rajković, A.,

Verlinden, B., Nicolaï, B., Debevere, J. and De Meulenaer, B. 2007. Influence

of storage conditions of apples on growth and patulin production by

Penicillium expansum. Int. J. Food Microbiol. 119:170–181.

Bayne, H. G., and Michener, H. D. 1979. Heat resistance of Byssochlamys ascospores.

Appl. Environ. Microbiol. 37:449–453.

Bennett, J. W., and Klich, M. 2003. Mycotoxins. Clin. Microbiol. Rev. 16:497–516.

Beuchat, L. R., and Pitt, J. I. 2001. Detection and enumeration of heat-resistant molds.

In Compendium of methods for the microbiological examination of foods, eds.

Frances P. Downes and Keith Ito. Washington, D.C.: The American Public

Health Assocation, p. 216–222.

Beuchat, L. R., and Rice, S. L. 1979. Byssochlamys spp. and their importance in

processed fruits. Adv. in Food Res. 25:237-288.

Beuchat, L. R., and Toledo, R. T. 1977. Behaviour of Byssochlamys nivea ascospores

in fruit syrups. Trans. Br. Mycol. Soc. 68:65–71.

Brown, A. H. S., and Smith, G. 1957. The genus Paecilomyces Bainier and its perfect

stage Byssochlamys Westling. Trans. Br. Mycol. Soc. 40:17–89.

Castresana J. 2000. Selection of conserved blocks from multiple alignments for their

use in phylogenetic analysis. Molecular Biology and Evolution. 17(4):540-552.

Darriba, D., Taboada, G. L., Doallo, R. and Posada, D. 2012. jModelTest 2: more

models, new heuristics and parallel computing. Nature methods. 9(2) 772-772.

Dijksterhuis, J. 2007. Heat-resistant ascospores. In Food Mycology: A Multifaceted

Approach to Fungi and Food., eds. Jan Dijksterhuis and Robert A. Samson.

CRC Press, p. 101–117.

Dombrink-Kurtzman, M. A., and Engberg, A. E. 2006. Byssochlamys nivea with

patulin-producing capability has an isoepoxydon dehydrogenase gene (idh)

with sequence homology to Penicillium expansum and P. griseofulvum. Mycol.

Res. 110:1111–8.

Edgar R.C. 2004. MUSCLE: multiple sequence alignment with high accuracy and

high throughput. Nucleic Acids Research 32(5):1792-1797

Geisler, B. M. 2013. “Commodity Apple Profile.” Agriculture Marketing Resource

Center. Available at: http://www.agmrc.org/commodities-

products/fruits/apples/commodity-apples/. Accessed: 21 January 2018.

Grant, W.D. 2004. Life at low water activity. Philos. Trans. R. Soc. London.

359:1249–1267.

Health Canada. 2018. “List of Contaminants and other Adulterating Substances in

Foods.” Available at: https://www.canada.ca/en/health-canada/services/food-

nutrition/food-safety/chemical-contaminants/contaminants-adulterating-

substances-foods.html#fn_t1b1-ref. Accessed: 21 January 2018.

Houbraken, J., and Samson, R. A. 2006. Standardization of methods for detecting heat

resistant fungi. In Advances in Food Mycology, eds. A. D. Hocking, J. I. Pitt,

R. A. Samson, and U. Thrane. New York, U.S.A.: Springer, p. 107–111.

Houbraken, J. and Samson, R. A. 2017. Current taxonomy and identification of

foodborne fungi. Current Opinion in Food Science. 17: 84-88.

Houbraken, J., Samson, R. A., and Frisvad, J. C. 2006. Byssochlamys: significance of

heat resistance and mycotoxin production. Adv. Exp. Med. Biol. 571:211–224.

Houbraken, J., Varga, J., Rico-Munoz, E., Johnson, S., and Samson, R. A. 2008.

Sexual reproduction as the cause of heat resistance in the food spoilage fungus

Byssochlamys spectabilis (anamorph Paecilomyces variotii). Applied and

Environmental Microbiology. 74(5): 1613-1619.

Huelsenbeck, J.P., Larget, B., and Alfaro, M.E. 2004. Bayesian phylogenetic model

selection using reversible jump Markov chain Monte Carlo. Molecular biology

and evolution. 21(6): 1123-1133.

Jurick, W. M., and Cox, K. D. 2016. Pre- and postharvest fungal apple diseases. In

Achieving Sustainable Cultivation of Apples, ed. Kate Evans. Burleigh Dodds

Science Publishing Limited, p. 1–12.

King, A. D., Michener, H. D., and Ito, K. A. 1969. Control of Byssochlamys and

related heat-resistant fungi in grape products. Appl. Microbiol. 18:166–173.

Li, W., Cowley, A., Uludag, M., Gur, T., McWilliam, H., Squizzato, S., Park, Y.M.,

Buso, N., and Lopez, R. 2015. The EMBL-EBI bioinformatics web and

programmatic tools framework. Nucleic Acids Research. 43:W580-W584.

Moake, M. M., Padilla-Zakour, O. I., and Worobo, R. W. 2005. Comprehensive

review of patulin control methods in foods. Compr. Rev. Food Sci. Food Saf.

4:8–21.

Paradis, E., Claude, J. and Strimmer, K. 2004. APE: analyses of phylogenetics and

evolution in R language. Bioinformatics. 20: 289–290.

Phillips, A. J. L., Crous, P. W., and Alves, A. 2007. Diplodia seriata, the anamorph of

“Botryosphaeria” obtusa. Fungal Divers. 25:141–155.

Piecková, E., Bernát, D., and Jesenská, Z. 1994. Heat resistant fungi isolated from soil.

Int. J. Food Microbiol. 22:297–299.

Puel, O., Galtier, P., and Oswald, I. P. 2010. Biosynthesis and toxicological effects of

patulin. Toxins (Basel). 2:613–31.

Ronquist, F., Teslenko, M., Van Der Mark, P., Ayres, D.L., Darling, A., Höhna, S.,

Larget, B., Liu, L., Suchard, M.A. and Huelsenbeck, J.P. 2012. MrBayes 3.2:

efficient Bayesian phylogenetic inference and model choice across a large

model space. Systematic Biology. 61(3): 539-542.

Salomão, B. D. C. M., Muller, C., Couto do Amparo, H., and Falcão de Aragão, G. M.

2014. Survey of molds, yeast and Alicyclobacillus spp. from a concentrated

apple juice productive process. Brazilian J. Microbiol. 45:49–58.

Samson, R. A., Hoekstra, E. S., Frisvad, J. C., and Filtenborg, O. 2000. Ascomycetes.

In Introduction to Food-and Airborne Fungi, eds. Robert A. Samson and Ellen

S. Hoekstra. Utrecht, The Netherlands: Centaalbureau voor Schimmelcultures,

p. 26–52.

Samson, R. A., Houbraken, J., Varga, J., and Frisvad, J. C. 2009. Polyphasic

taxonomy of the heat resistant ascomycete genus Byssochlamys and its

Paecilomyces anamorphs. Persoonia. 22:14–27.

Sant’Ana, A. D. S., Rosenthal, A., and de Massaguer, P. R. 2008. The fate of patulin

in apple juice processing: A review. Food Res. Int. 41:441–453.

Sanzani, S. M., Reverberi, M., Punelli, M., Ippolito, A, and Fanelli, C. 2012. Study on

the role of patulin on pathogenicity and virulence of Penicillium expansum. Int.

J. Food Microbiol. 153:323–31.

Schoch, C.L., Seifert, K.A., Huhndorf, S., Robert, V., Spouge, J.L., Levesque, C.A.,

Chen, W., Bolchacova, E., Voigt, K., Crous, P.W. and Miller, A.N. 2012.

Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA

barcode marker for Fungi. Proc. Natl. Acad. Sci. U. S. A. 109:6241–6.

Snini, S.P., Tadrist, S., Laffitte, J., Jamin, E.L., Oswald, I.P. and Puel, O. 2014. The

gene PatG involved in the biosynthesis pathway of patulin, a food-borne

mycotoxin, encodes a 6-methylsalicylic acid decarboxylase. International

Journal of Food Microbiology. 171:77-83.

Splittstoesser, D. F., Kuss, F. R., and Harrison, W. 1970. Enumeration of

Byssochlamys and other heat-resistant molds. Appl. Microbiol. 20:393–7.

Splittstoesser, D. F., Kuss, F. R., Harrison, W., and Prest, D. B. 1971. Incidence of

heat-resistant molds in eastern orchards and vineyards. Appl. Microbiol.

21:335–337.

Sutton, T. B., Aldwinckle, H. S., Agnello, A. M., and Walgenbach, J. F., eds. 2014.

Compendium of Apple and Pear Diseases and Pests. 2nd ed. St. Paul,

Minnesota U.S.A: The American Phytopathological Society.

Taniwaki, M. H., Hocking, A. D., Pitt, J. I., and Fleet, G. H. 2009. Growth and

mycotoxin production by food spoilage fungi under high carbon dioxide and

low oxygen atmospheres. Int. J. Food Microbiol. 132:100–108.

Taniwaki, M. H., Hocking, A. D., Pitt, J. I., and Fleet, G. H. 2010. Growth and

mycotoxin production by fungi in atmospheres containing 80% carbon dioxide

and 20% oxygen. Int. J. Food Microbiol. 143:218–225.

Tournas, V. 1994. Heat-resistant fungi of importance to the food and beverage

industry. Crit. Rev. Microbiol. 20:243–263.

US Apple Association. “Popular Varieties.” Available at: http://usapple.org/all-about-

apples/apple-varieties/. Accessed 21 January 2018.

US FDA. 2001. “Apple Juice, Apple Juice Concentrates, and Apple Juice Products -

Adulteration with Patulin.” Compliance policy guidance for FDA staff. Sec.

510.150. Available at:

https://www.fda.gov/ICECI/ComplianceManuals/CompliancePolicyGuidance

Manual/ucm074427.htm. Accessed: 21 January 2018.

Vaidya, G., Lohman, D.J., and Meier, R. 2011. SequenceMatrix: concatenation

software for the fast assembly of multi‐gene datasets with character set and

codon information. Cladistics. 27: 171-180.

Visagie, C.M., Houbraken, J., Frisvad, J.C., Hong, S.B., Klaassen, C.H.W., Perrone,

G., Seifert, K.A., Varga, J., Yaguchi, T. and Samson, R.A. 2014. Identification

and nomenclature of the genus Penicillium. Stud. Mycol. 78:343–371. van der Walt, L., Spotts, R. A., Visagie, C. M., Jacobs, K., Smit, F. J., and McLeod,

A. 2010. Species associated with preharvest Wet Core Rot in South Africa and

their pathogenicity on apple. Plant Dis. 94:666–675.

White, T. J., Bruns, S., Lee, S., and Taylor, J. 1990. Amplification and direct

sequencing of fungal ribosomal RNA genes for phylogenetics. In PCR

Protocols: A Guide to Methods and Applications, , p. 315–322.

Yates, A. R., Seaman, A., and Woodbine, M. 1968. Ascospore germination in

Byssochlamys nivea. Can. J. Microbiol. 14:319–325.

CHAPTER 3

SURVIVAL OF PAECILOMYCES NIVEUS IN APPLE JUICE CONCENTRATE

MADE FROM FRUIT WITH PAECILOMYCES ROT3

INTRODUCTION

Concerns and cost of heat-resistant mold spoilage.

In a recent survey, juice producers overwhelmingly identified spoilage as an important topic to the industry (36). Producers indicated that better control over microbial spoilage would increase profits and reduce waste (36). Respondents to the juice production survey (92 and 64%, respectively) reported experience with finished products that were contaminated with mold or yeast. However, surprisingly little research has been done on food spoilage compared to foodborne bacterial pathogens

(36). Heat-resistant mold (HRM) spoilage tends to be sporadic, making it difficult to study, and difficult to control.

Numerous factors contribute to HRM spoilage in finished products. These include the increasing demand from consumers for less-processed foods and fewer preservatives. The increasing prevalence of plastic containers that cannot obtain a true hermetic seal and cannot be held at high temperatures can lead to spoilage. In addition

3 Survival of Paecilomyces niveus in Apple Juice Concentrate Made from Fruit with Paecilomyces rot. 2018. Biango-Daniels, Megan N., Snyder, Abigail B., Worobo, Randy W., and Hodge, Kathie T. Submitted to the J Food Protection.

to the food waste and expensive recalls associated with spoilage, the damage to the reputation of a company in the globalized market can be hard to repair (36, 42).

Thermal processes designed with human pathogens in mind overlook HRM spoilage.

Specific time/temperature combinations vary among commercial thermal processes, which are designed primarily to control microorganisms that cause foodborne illnesses (22). Thermal processes for juice production are required by the

FDA to reduce vegetative pathogenic bacteria, such as E. coli O157:H7, and the parasite Cryptosporidium parvum by a minimum of 5-log (45). Established time- temperature combinations for low pH apple juice include a thermal pasteurization time of 3 sec at 71.1 °C, or 14 sec at 68.1 °C. Shelf-stable juices are processed at higher temperatures to control spoilage: One validated process is 2 sec at 90 °C, with a hot fill and hold at 85 °C for 1 min (45). During the production of apple juice concentrate a typical pasteurization is >92 °C for 30–60 sec, followed by thermal evaporation under vacuum to achieve a final Brix of approximately 70 (22). Vacuum evaporation is one of the most stringent thermal processes for apple products, yet there is evidence that HRM may survive (28). Thermal processing may inadvertently select for HRM, which survive and then face no competition for access to nutrients in products (22, 36,

45).

Paecilomyces niveus and associated HRM spoilage.

Paecilomyces niveus Stolk & Samson (Byssochlamys nivea Westling) is responsible for the spoilage of pasteurized foods from yogurts and fruit concentrates, to syrups and fruit sauces (7, 9, 40). If heat-resistant ascospores of P. niveus survive thermal processing, the fungus can grow in products stored across a wide range of temperatures (4-38°C) as a facultative anaerobe, even when there is as little as 0.5% oxygen in the head-space (37, 39). Paecilomyces niveus ascospores withstand temperatures greater than 85°C, irradiation, and high pressure (11, 13, 14, 15, 17, 35).

Little is known about the ecology of this species. It is common in orchard and vineyard soils and it has recently been shown to have the ability to infect apples, causing a post-harvest disease called Paecilomyces rot (6, 37).

Patulin is one of the most important chemical hazards in apple products (19).

Paecilomyces niveus is not the only patulin producing fungus associated with apples; the major contributor of patulin in apples is considered to be Penicillium expansum, a plant pathogen that causes the post-harvest disease called blue rot (18, 33). In addition to acute and subacute toxicity in animal models, there is evidence that patulin is carcinogenic (25). Patulin is one of only a few internationally regulated mycotoxins

(25). According to the U.S. Food and Drug Administration and Health Canada, the maximum allowable level of patulin in single strength juice is 50 ppb (µg L-1), while

EU has even lower maximum values for apple juices and products (10-25 ppb) (16,

44). The restrictive allowable levels were established based on the body weight and development stage of primary apple juice consumers, children, whose diets place them at high risk for exposure (4).

Patulin typically enters apple products through diseased fruit. It is controlled during juice production by culling apples and removing browned tissue. While P. expansum is killed by high temperatures during processing, Paecilomyces niveus survives and can produce patulin in juices after processing (26, 27). Previous work has demonstrated that when ascospores of P. niveus were introduced to apple juice packaged in polyethylene terephthalate bottles and laminated paperboard, patulin was produced in juice under shelf-storage conditions at 21 and 30 °C (32).

In contrast to P. expansum, Paecilomyces niveus infects and produces ascospores in apples that could be a source of contamination of apple products (6).

Fruit product contamination by P. niveus has traditionally been considered the result of unsanitary conditions and soil contamination, either of raw materials or during production (12, 17), however, the use of P. niveus infected fruit may introduce spoilage inoculum into thermally processed apple products, such as concentrates (6).

This study aimed to test the hypothesis that using apples infected with Paecilomyces rot could lead to apple juice concentrate contaminated with living P. niveus, which may produce patulin. In contrast to previous studies, which have evaluated the destruction of spores that were grown on plates and then added to juices, using infected apples as the route of contamination better represents the natural levels and physiological state of fungal contaminants. By sampling concentrate made with infected fruit at the various stages of processing, we assessed the levels of viable inoculum of P. niveus and measured patulin levels throughout the concentration process.

MATERIALS AND METHODS

Preparation of lab-infected apples.

To study the survival of the fungus throughout the concentration process,

‘Gala’ apples were artificially infected with a known isolate of P. niveus. ‘Gala’ was selected because of its availability and popularity as one the top ten commercial cultivars (43). Apples were surface-sterilized prior to inoculation with toothpick pieces

(16 mm) colonized by P. niveus isolate CO7 (6). Inoculum preparation and inoculation were as previously described (6, 46). Briefly, toothpick pieces were sterilized by repeated autoclaving in DI water, then transferred to potato dextrose broth (PDA,

Teknova) and autoclaved once more. Prepared toothpick pieces were then aseptically transferred to 100 mm Petri dishes containing PDA inoculated with three 5 mm plugs per plate of a 1-week-old CO7 culture. After 2 weeks incubation (30°C, dark), toothpick pieces were well colonized by P. niveus. To infect each apple, a colonized toothpick piece was inserted halfway between the stem and calyx. Inoculated ‘Gala’ apples were incubated in dark moist chambers (25°C, ≥ 95% humidity) for three weeks, allowing a significant lesion to develop (approximately 40 mm diameter) (6)

(Illustration 3.1).

Illustration 3.1. Representative images of a ‘Gala’ apple with (A) a lesion and (B) internal necrosis symptomatic of Paecilomyces rot (6).

Apple juice concentration.

Apple juice concentrate was made using a half peck (approximately 30 apples) of inoculated ‘Gala’ apples which were cider grade, with an average weight of 66.28 g

(n=3). The apples showed no signs or symptoms of diseases other than Paecilomyces rot (Illustration 3.1) (6). As a control, mock-inoculated apples were separately used to make concentrate. The latter apples were injured with toothpick pieces that were similarly prepared, except they were placed on fungus-free PDA Petri dishes before being used to wound apples. Samples taken from this batch of concentrate provided a baseline for the background quantity of P. niveus and the concentration of patulin normally occurring in the ‘Gala’ apples kept under the experimental conditions used in this study.

Apple juice concentration was performed at a small scale using methods typical of commercial practices (10) (Illustration 3.2). To prevent cross-contamination, mock-inoculated apples were processed prior to processing of inoculated apples.

Apples were cut and milled using a Waring blender, then hand-pressed through sterile cheesecloth to produce cider. Next, 50 mL tubes of cider were centrifuged (30 min,

3500 RPM), separating the apple solids from the juice. The juice was enzymatically clarified by adding concentrated pectinase (300g L-1, DSM Rapidase® C80 Max) and incubating for two hours at room temperature (25°± 2 C). Following incubation with the pectinase, the supernatant was decanted and briefly centrifuged in 50 mL tubes (5 min, 3500 RPM) to remove any remaining solids. Afterwards, 500 mL of the clarified juice (12 ± 2°Brix) was thermally processed by placing it in a rotating vacuum

(Rotovac) for 2.5 hours at 70 °C to produce apple juice concentrate (60 ± 2° Brix).

Brix were measured using a refractometer.

Illustration 3.2. The apple juice concentrate process used (left to right), illustrating the three sampling points after; (1) pressing (cider), (2) centrifugation (apple solids and juice), and (3) concentration (juice from concentrate).

Samples of cider or juice were taken throughout the process and visually examined using a microscope (Illustration 3.2) (Nikon Eclipse e600). These observations helped determine the nature of P. niveus inoculum, as intact asci and free ascospores. All liquid samples were observed using a compound microscope at 400X magnification, and propagules were counted with the aid of a hemocytometer. Apple solids were not examined, as the pulp prevented visualization of spores and asci.

Sampling and quantification of P. niveus across the concentrate process.

Samples were taken throughout the concentration process to quantify viable P. niveus. The product was sampled after pressing, centrifugation, and concentration

(Illustration 3.2). After centrifugation, apple solids were resuspended in sterile deionized (DI) water (5% w/v). Following pressing and centrifugation, 1 mL samples of juice were taken in duplicate and heat-treated (70°C, 30 min). Heat-treatment was used to reduce any background fungi present, which were not of interest to this study and increased the germination rate of P. niveus ascospores (37). At each sampling point (Illustration 3.2), serial dilutions were spread on potato dextrose agar (PDA:

Criterion). The apple juice concentrate was sampled after dilution with sterile DI water to produce 400 mL of 0.5x juice. This was added to equal parts 2x strength PDA and plated via a pour plate method to facilitate detection (24).

The plated samples of cider, juice, resuspended apples solids, and juice from concentrate were incubated for one month (dark, 25°C), during which time emerging colonies of P. niveus were counted and recorded (See Identification methods below).

For all plates, CFU were counted in the sample of highest dilution because at low dilutions, colonies of P. niveus merged and impeded counting and growth on pour plates was determined to result from a single surviving spore. The number of CFU L-1 was determined by multiplying CFU by the dilution factor, then dividing by the volume of the sample. Apple juice concentrate samples were enumerated by plate counting as well: A total of 49 pour plates on average were generated for each sample to accommodate the total volume of concentrate.

Fungal identification methods.

Paecilomyces niveus colonies that grew during incubation were identified by their growth rate and characteristic colony morphology (5, 30). For confirmation, one putative P. niveus colony per plate was subcultured on PDA and morphologically identified by microscopy of its spore forming structures using methods described by

Samson et al. (29).

A total of eight isolates were selected at random for verification of identities by sequencing: six isolates that were identified with microscopy and two isolates that were identified by colony morphology only. Mycelium was harvested from these eight

P. niveus cultures grown on PDA. Genomic DNA was extracted using PowerSoil

DNA isolation kit (MoBio). The extracted genomic DNA was stored in the elution buffer at -20 ̊ C. Using previously described methods for this species (6), the BenA region of the β-tubulin gene was amplified using forward Bt2a (5’-

GGTAACCAAATCGGTGCTGCTTTC -3’) and reverse Bt2b (5’-

ACCCTCAGTGTAGTGACCCTTGGC-3’) primers (48). This region is a widely accepted barcode for fungi was performed (34, 49). PCR amplicons were cleaned using the E.N.Z.A. MicroElute Cycle-pure kit (Omega Bio-tek) according to manufacturer instructions, and sequenced with the forward PCR primer, using an

Applied Biosystems Automated 3730xl DNA analyzer at the Cornell University

Biotechnology Resource Center (BRC). The isolates were identified by blasting

(Blastn) the sequences against NCBI fungal sequences (taxid: 4751) excluding uncultured/environmental sample sequences, and limited to sequences from type

material (1, 2). Blast results were supported with a phylogenetic analysis, using previously described methods (6).

Patulin analysis.

To determine if patulin was present, 20 mL frozen samples were taken at each of three points. Samples of cider (n=3) were taken from two experimental replicates made with inoculated apples and from the mock-inoculated control (Illustration 3.2).

One sample of juice, sampled after centrifugation, and one sample of single strength juice from concentrate were also analyzed (Illustration 3.2). Patulin analysis was performed by Trilogy Analytical Laboratory, Washington, M.O., U.S.A., utilizing

AOAC methods (23).

Statistical analysis.

To evaluate differences between mean CFU L-1 in each sample, a linear mixed effects model was used, with a fixed effect of sampling stage and a random effect of experimental replicate. The CFU counts were log transformed to meet the assumptions of linear models. An ANOVA was used to assess the model with the R statistical programming package lmerTest. Pairwise post-hoc comparisons were done using the lsmeans package.

RESULTS

Paecilomyces niveus survives concentration.

No P. niveus was isolated from any sample across the concentration process using control mock-inoculated apples. Viable P. niveus was detected in all samples taken across the juice concentrate process for lab-infected apples (Figure 3.1). All isolates from the subset taken from the dilution plates were morphologically identified as P. niveus. Blastn results of the β-tubulin region of the eight randomly selected P. niveus isolates that were sequenced (MG641435 - MG641442) supported their identification as P. niveus. The sequences were identical to those of reference isolates of P. niveus, including the type isolates and isolate CO7 (MG008852).

Cider made from infected apples contained an average of 5.5 log CFU L-1

(n=3, log standard deviation = 0.85) (Figure 3.1). After the first centrifugation step

(Illustration 3.2), the apple solids and juice contained an average of 4.3 log CFU L-1

(n=3, log standard deviation = 0.68) and 4.5 log CFU L-1 (n=2, log standard deviation

= 1.70), respectively (Figure 3.1). In one experimental replicate, the juice contained no detectable P. niveus. The juice from concentrate, when diluted to single strength, contained an average of 1.5 log CFU L-1 (n=3, log standard deviation = 0.26), which is

33.3 CFU L-1 (n=3, standard deviation = 16.1) (Figure 3.1). There were significantly less P. niveus in the single strength juice made from concentrate than in samples taken earlier in the process (p ≤ 0.005) (Figure 3.1). There was approximately a 4-log reduction in CFU L-1 between cider and juice from concentrate (Figure 3.1).

Microscopic examination revealed that samples of cider, juice, and juice from concentrate all contained ascospores that were morphology consistent with P. niveus.

The method used could distinguish asci and ascospores (Figure 3.2 A-B), and no intact asci were observed in any samples (Figure 3.2 C). The concentration process accounted for a 1.6 log reduction in observed spores (Figure 3.2 C).

Figure 3.1. Mean colony forming units per liter (CFU L-1) of P. niveus from samples taken throughout the apple juice concentrate process. Statistical significance is represented with * (P<0.05). Bars indicate standard deviation.

Figure 3.2. Representative images of P. niveus spore suspensions on a hemocytometer: (A) with more ascospores than asci (9:1); insert illustrates size and shape differences (not to scale), and (B) with mostly intact asci. The viability of these spores was not determined. The scale bars are 20 µm (A-B). The mean concentration of ascospores observed using a hemocytometer (C) at sampling points throughout the apple juice concentrate process.

Patulin concentration.

Cider sampled from the mock-inoculated concentrate provided a baseline for apples incubated in moist chambers. The cider from mock-inoculated apples contained

31 ppb patulin (Figure 3.3). Cider made from P. niveus-infected apples contained an average of 16,624 ppb patulin (11,677 and 21,572 ppb) (Figure 3.3). The juice made from infected apples, sampled after the first centrifugation (Illustration 3.2), contained

13,320 ppb, and single strength juice made from concentrate contained 19,409 ppb patulin (Figure 3.3). The uncertainty for patulin was 6.5%.

Figure 3.3. The patulin concentration (log ppb) in single strength cider made from mock-inoculated (control) apples (n=1), and the patulin concentration in single strength cider (n=2), juice (n=1), and juice from concentrate (n=1) made from apples that were inoculated with P. niveus. The horizontal line represents 50 ppb, the legal limit of patulin.

DISCUSSION

This is the first report that apples with Paecilomyces rot caused by Paecilomyces niveus can serve as a source of HRM contamination in apple juice concentrate.

Paecilomyces niveus spores decreased in abundance throughout the process, which

involved extensive heat processing, however, an average of 33.33 CFU L-1 were detected in the single strength juice from concentrate (Illustration 3.2). Our finding supports an old truism of food and beverage manufacturing: Quality of the end- product reflects the quality of the ingredients. This study demonstrates that using apples infected with P. niveus can result in apple juice concentrate contaminated with

HRM.

Paecilomyces species are the most abundant group of HRM in cider production

(38). Detected along apple juice concentrate production lines (28), their ascospores survive pasteurization and cause spoilage of bottled apple products (32). Prior research attributed the source of heat-resistant ascospores to soil contamination (12, 17).

However, since spores in infected fruit can survive pasteurization, fruit washing, filtration, and pasteurization, an important but previously unrecognized critical control step for P. niveus is the culling of fruit with Paecilomyces rot (Illustration 3.1).

Our results highlight that P. niveus infected apples introduce the “blind risk” of patulin, a mycotoxin and important chemical hazard for apple products (42).

Paecilomyces niveus has been recognized for its ability to make patulin in apple products in various packages on the shelf (32). The finding that patulin levels in two independent samples of cider made from apples with Paecilomyces rot were at least

200-times the maximum acceptable levels at the time of processing challenges the notion that this species primarily produces mycotoxins post-pasteurization (31)

(Figure 3.3). The patulin concentration in cider made from mock-inoculated apples was well below the allowed maximum 50 ppb established by the U.S. Food and Drug

Administration and Health Canada and was likely produced by other fungi (4, 16, 44).

Patulin-producing Penicillium species are common in apples where they can cause rots

(46) that are not always externally visible (Figure 5) (31). Patulin is heat stable and is not expected to decline during thermal processing (20, 31) (Figure 3.3). Additional and targeted evaluation and research is needed to determine the amount of patulin produced by P. niveus in apples. Future work should also quantify patulin accumulation in apples with Paecilomyces rot. Patulin is a virulence factor for blue rot disease, known to help increase disease (33), and whether the accumulation of patulin during Paecilomyces rot is also positively correlated with disease progression should be investigated.

Concentration is often a salvage activity, using apples rejected from the fresh market. These apples are of lesser quality due to long storage periods and they likely have defects (3, 28). Apple cider, and sometimes juice, contain spoilage fungi that originate on the fruit at harvest, in post-harvest environments, and during processing and packaging (41). Fallen apples, known as “drops,” are sometimes used for concentrate despite predisposition to rots and foodborne pathogen contamination via wounds (3, 8). Paecilomyces niveus could be transmitted on dropped fruit with adhering soil particles, or drops may be infected when wounds contact orchard soil that contains P. niveus (6, 17, 37).

Fewer CFU L-1 of P. niveus were counted than spores observed using microscopy (Illustration 3.2 and Figure 3.2), which may be explained by non-viable spores present in the product. Heat-resistant ascospores may not be viable for reasons including age. Alternatively, similarly sized chlamydospores or conidia may have been mistakenly included in visual counts: These spores are not known to survive heat

(30, 42). However, conidia are much less abundant in cultures (personal observations) and appear to be uncommon in infected apples (6, 30).

Paecilomyces niveus will continue to be a concern of juice manufacturers (36), given increasing consumer preference for less-processed foods with milder pasteurization (17, 47). Paecilomyces niveus spoilage may be prevented through appropriate storage and handling of apples for concentrate, namely culling apples with symptoms and excluding apples that had extended contact with soil. Future research may evaluate additional processing steps such as diatomaceous earth filtration (21), or different time-temperature combinations. More information on the frequency of P. niveus infection in apples and its ability to produce patulin are needed to help reduce risk associated with this heat resistant mold.

ACKNOWLEDGMENTS

We are grateful for Cornell University Orchard’s support of this work, including their donation of apples for our experiments. We thank Trilogy Analytical Laboratory,

Washington, MO, USA for patulin analysis, and Erika Mudrak of the Cornell

Statistical Consulting Unit for her data analysis advice and expertise. This work was supported by the USDA National Institute of Food and Agriculture, Hatch project

1002546 and by the Arthur Boller Apple Research Grant. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the National Institute of Food and

Agriculture (NIFA) or the United States Department of Agriculture (USDA).

REFERENCES

1. Agostino, M. 2012. Introduction to the BLAST suite and BLASTN, p. 47–71. In

Practical Bioinformatics, Garland Science Taylor and Francis, New York, NY.

2. Altschul, S.F., W. Gish, W. Miller, E.W. Myers, and D.J. Lipman. 1990. Basic

local alignment search tool. J. Mol. Biol. 215:403–410. doi:10.1016/S0022-

2836(05)80360-2.

3. Bates, R.P., J.R. Morris, and P.G. Crandall. 2001. Tree fruit: apple, pear, peach,

apricot and plums, p.151–171. In FAO Agricultural Services Bulletin 146:

Principles and Practices of Small- and Medium -Scale Fruit Juice Processing, Food

and Agriculture Organization of the United Nations, Rome.

4. Bennett, J.W., and M. Klich. 2003. Mycotoxins. Clin. Microbiol. Rev. 16(3):497–

516. doi:10.1128/CMR.16.3.497.

5. Beuchat, L.R., and J.I. Pitt. 2001. Detection and enumeration of heat-resistant

molds, p. 151–171. In F.P. Downes, and K. Ito (ed.), Compendium of Methods for

the Microbiological Examination of Foods, The American Public Health

Association, Washington, DC.

6. Biango-Daniels, M.N., and K.T. Hodge. 2018. Paecilomyces rot: A New Apple

Disease. Pending publication Plant Dis. https://doi.org/10.1094/PDIS-12-17-1896-

RE.

7. Brown, A.H.S., and G. Smith. 1957. The genus Paecilomyces Brainier and its

perfect stage Byssochlamys Westling. Trans. Br. Mycol. Soc. 40(1):17–89.

doi:10.1016/S0007-1536(57)80066-7.

8. Cai, S., R.W. Worobo, and A.B. Snyder. 2018. Outgraded produce variably retains

surface inoculated Escherichia coli through washing. Pending publication Int. J. of

Food Micro. 269:27-35.

9. Dagnas, S., and J. M. Membré. 2013. Predicting and preventing mold spoilage of

food products. J. of Food Prot. 76(3): 538-551. doi:10.4315/0362-028X.JFP-12-

349.

10. Downing, D.L. 1989. Processed Apple Products. Springer, New York, NY.

11. Evelyn, E. and F.V.M. Silva. 2015. Inactivation of Byssochlamys nivea ascospores

in strawberry puree by high pressure, power ultrasound and thermal processing.

Int. J. Food Microbiol. 214:129–136. doi:10.1016/j.ijfoodmicro.2015.07.031.

12. Frąc, M., S. Jezierska-Tys, and T. Yaguchi. 2015. Occurrence, detection, and

molecular and metabolic characterization of heat-resistant fungi in soils and plants

and their risk to human health. Adv. Agron. 132:161–204.

doi:10.1016/bs.agron.2015.02.003.

13. Gumus, T., U. Gecgel, A.S. Demirci, and M. Arici. 2008. Effects of gamma

irradiation on two heat resistant moulds: Aspergillus fumigatus and Paecilomyces

variotii isolated from margarine. Radiat. Phys. Chem. 77:680–683.

doi:10.1016/j.radphyschem.2007.10.001.

14. Hamanaka, D., G.G. Atungulu, F. Tanaka, and T. Uchino. 2010. Effect of

combining infrared heating with ultraviolet irradiation on inactivation of mold

spores. Food Sci. Technol. Res. 16(4):279–284. doi:10.3136/fstr.16.279.

15. Hatcher, W.S., J.L. Weihe, D. I. Murdock, J.F. Folinazzo, E.C. Hill, and L.G.

Albrigo. 1979. Growth requirements and thermal resistance of fungi belonging to

the genus Byssochlamys. J. Food Sci. 44(1):2–6. doi: 10.1111/j.1365-

2621.1979.tb10021.x.

16. Health Canada. 2016. Maximum levels for chemical contaminants in foods.

Available at: https://www.canada.ca/en/health-canada/services/food-

nutrition/food-safety/chemical-contaminants/maximum-levels-chemical-

contaminants-foods.html. Accessed 29 January 2018.

17. Houbraken, J., J. Dijksterhuis, and R.A. Samson. 2012. Diversity and biology of

heat-resistant fungi, p. 331–353. In Wong, H. (ed.), Stress Response of Foodborne

Microorganisms, Nova Science Publishers Inc., Hauppauge, NY.

18. Houbraken, J., and R.A. Samson. 2006. Standardization of methods for detecting

heat resistant fungi, p 107-111. In Hocking, A.D., J.I. Pitt, J.I., R.A. Samson, and

U. Thrane (ed.), Advances in Food Mycology. Springer, New York, NY.

19. Houbraken, J., R. A. Samson, J.C. Frisvad. 2006. Byssochlamys: significance of

heat resistance and mycotoxin production. Adv. Exp. Med. Biol. 571:211–224.

doi:10.1007/0-387-28391-9_14.

20. Kadakal, C., and S. Nas. 2003. Effect of heat treatment and evaporation on patulin

and some other properties of apple juice. J. Sci. Food Agric. 83(9): 987-990.

doi:10.1002/jsfa.1339.

21. King, A. D., H.D. Michener, and K.A. Ito. 1969. Control of Byssochlamys and

related heat-resistant fungi in grape products. Appl. Microbiol. 18(2):166–173.

22. Lawlor, K. A., J.D. Schuman, P.G. Simpson, and P.J. Taormina. 2009.

Microbiological spoilage of beverages, p. 245-284. In Sperber, W. H. and M.P.

Doyle (ed.), Compendium of microbiological spoilage of foods and beverages,

Springer, New York, NY.

23. Official Methods of Analysis of AOAC INTERNATIONAL. 2005. 18th ed.,

AOAC INTERNATIONAL, Gaithersburg, MD, USA, Official Method 995.10.

24. Pitt, J.I., and A.D. Hocking. 1997. Methods for isolation, enumeration and

identification, p. 21–39. In Fungi and Food Spoilage, Blackie Academic and

Professional, New South Wales, Australia.

25. Puel, O., P. Galtier, and I.P. Oswald. 2010. Biosynthesis and toxicological effects

of patulin. Toxins. 2(4):613–31. doi:10.3390/toxins2040613.

26. Rice, S.L., L.R. Beuchat, and R.E. Worthington. 1977. Patulin production by

Byssochlamys spp. in fruit juices. Appl. Environ. Microbiol. 34(6):791-796

27. Roland, J.O., and L.R. Beuchat. 1984. Biomass and patulin production by

Byssochlamys nivea in apple juice as affected by sorbate, benzoate, SO2 and

Temperature. J. Food Sci. 49(2): 402–406. doi: 10.1111/j.1365-

2621.1984.tb12432.x.

28. Salomão, B.C.M., C. Muller, H. Couto do Amparo, and G.M. Falcão de Aragão.

2014. Survey of molds, yeast and Alicyclobacillus spp. from a concentrated apple

juice productive process. Brazilian J. Microbiol. 45(1):49–58. doi:

10.1590/S1517-83822014000100008.

29. Samson, R.A., E.S. Hoekstra, J.C. Frisvad, and O. Filtenborg. 2000. Identification

of the common food- and airborne fungi, p. 28–30. In Samson, R.A. and E.S.

Hoekstra. (ed.), Introduction to Food-and Airborne Fungi, CBS, Utrecht, The

Netherlands.

30. Samson, R.A., J. Houbraken, J. Varga, and J.C. Frisvad. 2009. Polyphasic

taxonomy of the heat resistant ascomycete genus Byssochlamys and its

Paecilomyces anamorphs. Persoonia, 22: 14–27.

doi:10.3767/003158509X418925.

31. Sant’Ana, A.D.S., A. Rosenthal, and P.R. Massaguer. 2008. The fate of patulin in

apple juice processing: A review. Food Res. Int., 41(5): 441-453. doi:

10.1016/j.foodres.2008.03.001.

32. Sant'Ana, A.D.S., R.C. Simas, C.A. Almeida, E.C. Cabral, R.H. Rauber, C.A.

Mallmann, M.N. Eberlin, A. Rosenthal, A., and P.R. Massaguer. 2010. Influence

of package, type of apple juice and temperature on the production of patulin by

Byssochlamys nivea and Byssochlamys fulva. Int. J. Food Microbiol., 142:156–63.

doi:10.1016/j.ijfoodmicro.2010.06.017.

33. Sanzani, S.M., M. Reverberi, M. Punelli, A. Ippolito, and C. Fanelli. 2012. Study

on the role of patulin on pathogenicity and virulence of Penicillium expansum. Int.

J. Food Microbiol., 153(3):323–331. doi:10.1016/j.ijfoodmicro.2011.11.021.

34. Schoch, C.L., K.A. Seifert, S. Huhndorf, V. Robert, J.L. Spouge, C.A. Levesque,

and W. Chen. 2012. Nuclear ribosomal internal transcribed spacer (ITS) region as

a universal DNA barcode marker for Fungi. Proc. Natl. Acad. Sci. 109(16):6241–

6246. doi:10.1073/pnas.1117018109.

35. Smelt, J.P.P.M., and S. Brul. 2014. Thermal inactivation of microorganisms. Crit.

Rev. Food Sci. and nutrition. 54(10):1371–85.

doi:10.1080/10408398.2011.637645.

36. Snyder, A.B., and R.W. Worobo. 2018. The incidence and impact of microbial

spoilage in the production of fruit and vegetable juices as reported by juice

manufacturers. Food Control, 85:144–150. doi:10.1016/j.foodcont.2017.09.025.

37. Splittstoesser, D.F., F.R. Kuss, W. Harrison, and D.B. Prest. 1971. Incidence of

heat-resistant molds in eastern orchards and vineyards. Appl. Microbiol.

21(2):335–337.

38. Swanson, K.M.J., S.B. Leasor, and D.L. Downing. 1985. Aciduric and heat

resistant microorganisms in apple juice and cider processing operations. J. Food

Sci. 50:336–339. doi: 10.1111/j.1365-2621.1985.tb13395.x.

39. Taniwaki, M.H., A.D. Hocking, J.I. Pitt, and G.H. Fleet. 2009. Growth and

mycotoxin production by food spoilage fungi under high carbon dioxide and low

oxygen atmospheres. Int. J. Food Microbiol. 132(2):100–108.

doi:10.1016/j.ijfoodmicro.2009.04.005.

40. Tournas, V. 1994. Heat-resistant fungi of importance to the food and beverage

industry. Crit. Rev. Microbiol. 20(4):243–263. doi:10.3109/10408419409113558.

41. Tournas, V.H., J. Heeres, and L. Burgess. 2006. Moulds and yeasts in fruit salads

and fruit juices. Food Microbiol. 23(7):684–688. doi:10.1016/j.fm.2006.01.003.

42. Tribst, A.A.L., A.D.S. Sant’Ana, and P.R. de Massaguer. 2009. Review:

Microbiological quality and safety of fruit juices—past, present and future

perspectives. Crit. Rev. Microbiol. 35(4):310-339.

doi:10.3109/10408410903241428.

43. US Apple Association. Popular Varieties. Available at: http://usapple.org/all-

about-apples/apple-varieties/. Accessed 29 January 2018.

44. US FDA. 2001. CPG Sec.510.150. Apple juice, apple juice concentrates, and apple

juice products - adulteration with patulin. Available at:

https://www.fda.gov/ICECI/ComplianceManuals/CompliancePolicyGuidanceMan

ual/ucm074427.htm. Accessed 5 April 2018.

45. US FDA. 2004. Guidance for Industry: Juice HACCP Hazards and Controls

Guidance First Edition. Available at:

https://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryIn

formation/Juice/ucm072557.htm#vii. Accessed 29 January 2018.

46. van der Walt, L., R.A. Spotts, C.M. Visagie, K. Jacobs, F.J. Smit, and A McLeod.

2010. Species associated with preharvest wet core rot in South Africa and their

pathogenicity on apple. Plant Dis. 94(6):666–675. doi:10.1094/PDIS-94-6-0666.

47. Vermeulen, A., P. Ragaert, A. Rajkovic, S. Samapundo, F. Lopez-Galvez, and F.

Devlieghere. 2013. New research on modified-atmosphere packaging and

pathogen behavior, p. 340–354. In Advances in Microbial Food Safety, Woodhead

Publishing Series in Food Science, Technology and Nutrition, Cambridge, UK.

48. Visagie, C.M., J. Houbraken, J.C. Frisvad, S.B. Hong, C.H.W. Klaassen, G.

Perrone, K.A. Seifert, J. Varga, T. Yaguchi, and R.A. Samson. 2014. Identification

and nomenclature of the genus Penicillium. Stud. Mycol. 78:343–371.

doi:10.1016/j.simyco.2014.09.001.

49. White, T.J., S. Bruns, S. Lee, and J. Taylor. 1990. Amplification and direct

sequencing of fungal ribosomal RNA genes for phylogenetics, p. 315–322. In PCR

Protocols: A Guide to Methods and Applications, Academic Press, Cambridge,

MA.

CHAPTER 4

PAECILOMYCES NIVEUS: PATHOGENICITY IN THE ORCHARD AND

SENSITIVITY TO THREE FUNGICIDES4

INTRODUCTION

The United States is the world’s second-largest apple producer. This economically important crop has an estimated annual wholesale value of $4 billion (US Apple

Association). Commercial apples producers face numerous fungal diseases, especially during post-harvest, when apples are stored for periods exceeding 6 months (Sutton et al. 2014). While average annual losses of apples during post-harvest storage in the

United States are on the order of 10-15%, losses are typically greater in markets in developing countries (Jurick and Cox 2016).

Paecilomyces niveus Stolk & Samson (Byssochlamys nivea Westling)

(Ascomycota, Eurotiomycetes, Eurotiales, Thermoascaceae) has long been recognized as an important heat-resistant mold implicated in spoilage (Houbraken et al. 2006;

Houbraken and Samson 2017; Tournas 1994). Along with its relatives, P. variotti (B. spectabilis) and P. fulvus (B. fulva) (Samson et al. 2009), Paecilomyces niveus is one

4 Biango-Daniels, Megan N., Ayer, Katrin M., Cox, Kerik D., and Hodge, Kathie T. 2018. Paecilomyces niveus: Pathogenicity in the orchard and sensitivity to three fungicides. Submitted to J Plant Disease.

of the most difficult heat-resistant molds to inactivate during thermal processes such as pasteurization (Bayne and Michener 1979; Tournas 1994). Paecilomyces species are the most abundant heat-resistant molds in apple cider production (Swanson et al.

1985). Ascospores of P. niveus survive temperatures of ≥85°C, enabling it to spoil processed foods that range from fruit concentrate to syrup (Brown and Smith 1957;

Tournas 1994). Recently this species has been described as a post-harvest pathogen of apples, causing Paecilomyces rot (Biango-Daniels and Hodge 2018).

Paecilomyces rot resembles other common apple diseases, which may make it difficult to detect. External symptoms include brown circular flattened lesions, with faint concentric rings that resemble black rot or bull’s-eye rot (Biango-Daniels and

Hodge 2018; Sutton et al. 2014). The internal necrosis caused by Paecilomyces rot is bruise-like. It most closely resembles bitter rot disease, but the tissue is considerably firmer (Biango-Daniels and Hodge 2018; Phillips et al. 2007; Sutton et al. 2014).

Although little is known about the ecology of P. niveus, it has been isolated from both orchard and vineyard soils (Biango-Daniels and Hodge 2018; King et al. 1969).

Paecilomyces niveus is of additional concern because it can produce patulin, the most hazardous mycotoxin associated with apples (Houbraken et al. 2006).

Patulin, which is also produced by the post-harvest apple pathogen, Penicillium expansum, is one of only a few internationally regulated mycotoxins (Houbraken and

Samson 2006; Sanzani et al. 2012). It is limited to 50 ppb by the U.S. Food and Drug

Administration and Health Canada (Health Canada 2016; US FDA 2001), and 10 ppb by the EU (Puel et al. 2010). Paecilomyces niveus differs from other patulin-producing

species in its ability to produce patulin in packaged apple juice under shelf-storage conditions (Rice et al. 1977; Roland and Beuchat 1984; Sant’Ana et al. 2010).

Fungicides are one of the few tools available to prevent or mitigate these losses caused by post-harvest fungal pathogens. Three common fungicides used to control post-harvest diseases are difenoconazole, fludioxonil, and pyrimethanil, which represent FRAC groups 3, 9, and 12 respectively, as defined by the Fungicide

Resistance Action Committee (FRAC) (FRAC 2018). Difenoconazole, fludioxonil, and pyrimethanil are found in the commercial products Academy (Syngenta, Basel,

Switzerland), Scholar & Academy (Syngenta), and Scala (Bayer CropScience;

Monheim am Rhein, Germany) respectively. Some fungicides in these FRAC groups are also used early in the apple growing season to manage diseases like apple scab and powdery mildew, in the summer to manage diseases like sooty blotch and flyspeck, and as post-harvest treatments against rot diseases (Cox 2017). While fungicides can be highly effective in controlling fungal diseases of apple, reductions in sensitivity or complete resistance has limited their overall efficacy, creating problems for management (Lesniak et al. 2011; Villani et al. 2015).

The published description of Paecilomyces rot on apples has led some to question the traditional wisdom that P. niveus is introduced solely through soil contamination or unsanitary conditions during production (Brown and Smith 1957;

Biango-Daniels and Hodge 2018); could infected apples also be a source of contamination? Critical research is needed to understand the etiology of Paecilomyces rot in the orchard and the activity of fungicides against P. niveus, related to the risk of resistance development. This study aimed to evaluate the ability of P. niveus to infect

apples in orchards, and to characterize the sensitivity of P. niveus isolates to three different fungicide active ingredients; pyrimethanil, difenoconazole, and fludioxonil.

Specific objectives were to: (1) determine pathogenicity of P. niveus in an apple orchard, (2) compare any symptoms resulting from inoculation in an orchard to symptoms of post-harvest Paecilomyces rot, (3) determine the in vitro efficacy of post- harvest fungicides against P. niveus, and (4) compare fungicide sensitivity of baseline and exposed P. niveus isolates.

MATERIALS AND METHODS

Identification of P. niveus isolates.

To identify all P. niveus isolates used in the fungicide study (Table 4.1), as well as the fungi reisolated from infected apples in the orchard disease trial, pure cultures were first morphologically identified using the Ascomycete key by Samson et al. (2000). To confirm morphological identities, and distinguish P. niveus from close relatives, the genomic DNA of each isolate was extracted with the PowerSoil DNA

Isolation Kit (MoBio) following the recommended protocol, and the β-tubulin region

(benA gene) was subject to PCR amplification as previously described (Biango-

Daniels and Hodge 2018; Visagie et al. 2014). Resulting amplicons were sequenced with the forward PCR primer, using an Applied Biosystems Automated 3730xl DNA analyzer, at the Cornell University Biotechnology Resource Center and accession numbers were deposited in NCBI (Table 4.1).

Table 4.1. Origins of exposed and baseline P. niveus isolates used to characterize fungicide sensitivity and NCBI accession numbers for benA region DNA sequence.

NCBI Group Isolate Origin Accession Exposed 100-10 Orchard soil, Ithaca, NY MG008853 Exposed 101-3 Orchard soil, Geneva, NY MG008854 Exposed 102-1 Orchard soil, Geneva, NY MG008855 Exposed 104-22 Orchard soil, Peru, NY MG008856 Exposed 106-3 Orchard soil, Peru, NY MG008857 Exposed 108-11 Orchard soil, Peru, NY MG008859 Exposed 109-21 Orchard soil, Peru, NY MG008860 Exposed 110-3 Orchard soil, Peru, NY MG008861 Exposed 111-1 Orchard soil, Peru, NY MG008862 Exposed 112-22 Orchard soil, Chazy, NY MG008863 Exposed 125-31 Orchard soil, Sodus, NY MG008864 Exposed 140-11 Orchard soil, Shenandoah, NY MG008865 Exposed 141-11 Orchard soil, New Paltz, NY MG008866 Exposed 146-311 Orchard soil, Newark, NY MG008867 Exposed AF01 Alfalfa field soil, Owego NY MH046877 Exposed CF4 Corn field soil, Berkshire, NY MH046873 Exposed CFO3 Corn field soil, Owego, NY MH046874 Exposed CO7 Cornell Orchards apple, Ithaca NY MG008852 Exposed KRA4 Processed food, Geneva, NY MH046875 Exposed STE Decaying ‘Gala’ Apple, Ithaca, NY MH046876 CC8, MH046878 Baseline Cornell compost soil, Ithaca, NY CC12 MH046882 MH046879 MC8, MC5, Baseline Garden soil, Ithaca, NY MH046880 MC4 MH046881 MH046883 Baseline EP1, EP3 Wetland soil, Seneca Falls, NY MH046884 MH046885 BY8, BY3, Baseline Barnyard soil, Speedsville, NY MH046886 BY11 MH046887

Morphological identities of isolates confirmed using benA sequences of isolates in this study were compared using phylogenetic analysis to well-studied isolates in the same genus (Samson et al. 2009). Accession numbers for reference

sequences, provided in Samson et al. (2009), were used to acquire sequences for benA from NCBI GenBank. Next, the sequences from putative P. niveus isolates generated in this study were aligned to those used in Samson et al. (2009) using MUSCLE

(Edgar, 2004) via the EMBL-EBI online bioinformatics resource (Li et al. 2015), with poorly aligned, divergent regions trimmed from the alignment using GBlocks version

0.91b (Castresana 2000), specifying the less stringent criteria. The total number of taxa and characters in the partial β-tubulin analysis were 67 and 429 respectively.

A tree was constructed by Bayesian inference using MrBayes v3.2.6 (Ronquist et al. 2012). The partial β-tubulin analysis was conducted using a mixed model.

Posterior probabilities at branch nodes were inferred using two runs of 1,000,000 generations, with 25% burn-in. Phylogenetic trees were visualized in R v3.2.3, using the Ape package (Paradis et al. 2004), in the RStudio v1.1.383 environment, rooted with Thermoascus crustaceus as the outgroup.

Preparation of P. niveus inoculum.

Inoculum on colonized toothpick pieces was prepared as previously described for characterization of P. niveus post-harvest infection and for Penicillium species that cause wet core rot in apples (Biango-Daniels and Hodge 2018; van der Walt et al.

2010). Paecilomyces niveus CO7 (NRRL 66824), an isolate previously used to study post-harvest Paecilomyces rot which was originally isolated from a decaying apple at

Cornell Orchards, Ithaca, New York was used to generate all inoculum (Biango-

Daniels and Hodge 2018).

In short, inoculated toothpick pieces (16 mm) were first sterilized and incubated on PDA Petri dishes (100 mm) with an agar plug (5 mm) of a one-week old

P. niveus CO7 culture (Biango-Daniels and Hodge 2018). After two weeks of incubation (30°C, dark) toothpicks were completely colonized by P. niveus. Toothpick pieces used for the mock-inoculation were treated the same, however they were placed on PDA Petri dishes without the fungus.

Inoculation of apples in the orchard.

Apples were inoculated on trees in a 17-year-old orchard planting of ‘Gala’ trees on B.9 rootstocks at the New York State Agricultural Experiment Station in

Geneva, New York. The orchards were planted as a randomized block design with four replicate blocks for each plot designated for treatment. In July, apples were inoculated with P. niveus (n=55) or mock-inoculated with an uncolonized toothpick piece (n=42). One toothpick piece was inserted approximately halfway between the stem and calyx of immature apples. Inoculated and mock-inoculated apples were marked with different colored flags, tied near stems. At the time of inoculation apples weighed an average of 44.7 g ± 11.85 standard deviation (sd) (n=34) and had no visible defects or diseases.

Measurement of P. niveus symptoms in the orchard.

After one month, a subsample of apples (Treatment n=5, Control n=4) were harvested from across the four replicate plots to assess the efficacy of the inoculation method and progression of infection. Observations of lesion size were made and fungi

were isolated from several inoculated apples for identification. The remaining apples were harvested after eight weeks, giving lesions time to fully develop.

Immediately following harvest, apples were refrigerated for 48 hours, and subsequently evaluated for symptoms and apple development in terms of lesion diameter, apple weight, and roundness. Lesion diameter was measured at two perpendicular angles, using digital calipers (VWR Carbon Fiber Composite), and averaged. Roundness of apples was measured as the difference between the diameter of the uninoculated axis and inoculated axis (Illustration 4.1). To confirm that P. niveus was present in infected apples, fungal isolations were made from lesions on inoculated apples. The fungi isolated were grown in pure culture and identified (See section: Identification of P. niveus isolates).

To test the effect of P. niveus inoculation on three apple properties (lesion diameter, weight, and roundness) ANOVA models were used that included fixed effects of orchard block, treatment, and their interaction. Lesion diameter data were log-transformed to meet the assumptions of normality. The models were evaluated in the R statistical software with post-hoc analyses using the lsmeans package. The data were visualized using ggplot2 package.

Illustration 4.1. A represented image of an inoculated ‘Gala’ apple, illustrating how roundness was measured; as the difference between the diameter of the uninoculated axis (a) and inoculated axis (b).

Characterization of fungicide sensitivity in baseline isolates of P. niveus.

The baseline sensitivity of P. niveus isolates to post-harvest fungicides containing difenoconazole, fludioxonil, and pyrimethanil was determined using for ten

P. niveus isolates that originated from non-agricultural environments (Table 4.1).

Being collected from non-agricultural settings, these isolates were unlikely to have been exposed to agricultural fungicides. Difenoconazole, fludioxonil, and pyrimethanil were chosen because they represent different FRAC groups and are frequently used to manage post-harvest diseases on apple.

The effective concentration needed to inhibit mycelial growth by 50% (EC50) was determined for baseline P. niveus isolates using PDA amended with commercial formulations of difenoconazole (Inspire; Syngenta Agrochemical Company), fludioxonil (Scholar; Syngenta Agrochemical Company), and pyrimethanil (Scala;

Bayer Crop Science) dissolved in distilled water. Plugs (3mm) of P. niveus isolates were transferred from the edge of actively growing 3-week old cultures (dark, 25 ± 2

°C) to PDA plates that had been amended with each fungicide at the following doses:

0 (control), 0.001, 0.01. 0.05, 0.1, 0.5. 1.0, 5.0, and 10 μg ml-1. Plates were incubated at room temperature (24± 2 °C). The mycelial growth on each fungicide-amended plate was measured after one week with two replicate colonies per isolate.

The percent relative growth was calculated for each fungicide concentration as previously described (Frederick et al. 2015; Villani et al. 2015). A probit analysis was performed using the Statistical Analysis System (SAS) Version 9.4 (Cary, NC) to calculate the effective concentration needed for 50% inhibition based on the PDA control as previously described (Kreis et al. 2016).

Comparisons of fungicide sensitivity between baseline and exposed P. niveus isolates.

Twenty P. niveus isolates that originated from environments where fungicides were likely used were designated as exposed isolates (Table 4.1); ten others from non- agricultural sources were designated as baseline isolates. Some exposed P. niveus isolates originated from production soils other than apple as previously described

(Biango-Daniels and Hodge 2018).

To allow for direct comparisons of fungicide sensitivity between baseline and exposed isolates, isolates were subjected to a relative growth assay using discriminatory doses based on the mean EC50 value for baseline isolates. Plugs (3mm) of exposed P. niveus isolates were transferred from the edge of actively growing 3-

week old cultures (dark, 25 ± 2 °C) to PDA plates that had been amended with the

-1 approximate EC50 concentrations of each fungicide (difenoconazole 0.1 μg ml , fludioxonil 0.01 μg ml-1, and pyrimethanil 0.5 μg ml-1), as well as a control PDA plate, which had no added fungicide. Plates were incubated, diameter of growth was measured, and percent relative growth was calculated for two replicate colonies for each isolate as described above.

To visualize the relationship among exposed and baseline isolates of P. niveus, the percent relative growth data at the EC50 concentrations of fungicides were used to construct a heat map. The R-package heatmap.2 was used with the package defaults, including hierarchical clustering algorithm and Euclidean distances. A linear mixed effects model was used to compare percent relative growth of baseline and exposed P. niveus isolates on difenoconazole, fludioxonil, and pyrimethanil. This model included fixed effects of fungicide active ingredient and group (exposed and baseline), and a random effect of P. niveus isolate. The response was percent relative growth after one week. Post-hoc analyses were done with Tukey's honest significance test to determine pairwise differences among treatment combinations. This model was evaluated in R v3.2.3 using the lme4 and lmerTest packages. The post-hoc analyses were done using the lsmeans package and data were visualized using ggplot2 package.

RESULTS

Identification of P. niveus isolates.

Paecilomyces niveus that were isolated from inoculated apples at the mid- and endpoint of the orchard experiment were grown in pure cultures, all of which were morphologically identified as P. niveus. DNA sequences of the β-tubulin (benA) gene were deposited in NCBI (MH04688-MH046893). The morphological identities of these isolates (isolates A-F in Figure 4.1), as well as all isolates used in the fungicide sensitivity assays (Table 4.1), were supported by phylogenetic analysis, in which all isolates clustered with well-studied P. niveus (B. nivea) isolates (Samson et al., 2009)

(Figure 4.1).

Characteristics of symptoms caused by P. niveus in the orchard.

After eight weeks, the average lesios diameters on mock-inoculated (n=35) and inoculated (n=42) apples were 3.34 mm ± 1.85 sd, and 11.17 mm ± 6.82 sd, respectively. The inoculated apples in all four experimental blocks had lesions that were significantly larger than mock-inoculated lesions (P<0.005) (Figure 4.2).

Lesions on inoculated apples were brown and circular with faint concentric rings of alternating browns (Illustration 4.2 A). The center of the lesion was a lighter brown, which progressively darkened to the margins, which were dark-brown and diffuse (Illustration 4.2 A). No external sporulation was observed. The internal necrotic symptoms included a medium-brown, firm, and U-shaped rot (Illustration 4.2

B). In the mesocarp of inoculated apples a callus formed at the interface of necrotic

and healthy tissue that compartmentalized the rot (Illustration 4.3 A). When the rot extended to the core, white mycelia and asci were produced in the air space around the seeds (Illustration 4.3 B).

Figure 4.1. Phylogenetic relationships of isolates from this study plus related species. Consensus topology and branch lengths of partial β-tubulin sequences of 67 Paecilomyces (Byssochlamys) taxa, including isolates used in fungicide susceptibility assays and reisolated from inoculated apples (this study, in bold), rooted with Thermoascus crustaceus as an outgroup. Tree constructed with MrBayes v3.2.6. Support values at nodes represent posterior probabilities, scale bar shows the inferred number of substitutions per site.

Figure 4.2. The diameter of inoculated and mock-inoculated lesions on ‘Gala’ apples at harvest (8 weeks post-inoculation). In each orchard block inoculated lesions were significantly larger than mock-inoculated lesions (P<0.005).

Illustration 4.2. Representative symptoms of Paecilomyces Rot on ‘Gala’ apples 8 weeks post-inoculation. In cross section, a firm brown U-shaped necrotic area developed beneath a flattened surface where the inoculated toothpicks were inserted, leading to notable growth asymmetry (A). A brown lesion forming around the point of inoculation, with faint concentric rings and diffuse margins (B).

Illustration 4.3. Representative images of the mesocarp of inoculated apples with distinct callus development at the edge of the lesion, scale bar = 1 mm (A). Asci and ascospores of Paecilomyces niveus (Byssochlamys nivea) produced inside the core of an inoculated apple, scale bar = 10µm (B).

At harvest there was no significant weight difference between inoculated and mock-inoculated apples (P=0.229). The average weight of mock-inoculated and inoculated apples was 134 g ± 55.6 sd and 133.6 g ± 47.1 sd respectively. There was a significant difference in roundness between mock-inoculated and inoculated apples

(p=0.009). Inoculated apples were visibly flattened on the face of the apple that had been wounded, giving them an asymmetrical, lopsided shape (Illustration 4.1,

Illustration 4.3).

Sensitivity of P. niveus isolates to fungicides.

To assess current or future shifts towards insensitivity in exposed populations, baseline sensitivity of P. niveus was determined for difenoconazole, fludioxonil, and pyrimethanil. The EC50 values for baseline P. niveus isolates to difenoconazole,

100

fludioxonil, and pyrimethanil were variable. They ranged from 0.0137–0.593 μg/ml-1,

0.000587–0.02789 μg/ml-1, and 0.261–0.868 μg/ml-1 for difenoconazole, fludioxonil, and pyrimethanil respectively. Fludioxonil had the highest level of activity against baseline isolates of P. niveus. The mean EC50 for fludioxonil was determined to be

0.015 μg/ml-1. Pyrimethanil had the lowest activity against baseline isolates, with a

-1 mean EC50 of 0.464 μg/ml . Difenoconazole had intermediate activity against baseline

-1 isolates and the mean EC50 was 0.194 μg/ml .

Figure 4.3. The percent relative growth of baseline and exposed P. niveus isolates after one week at adjusted mean EC50 concentrations of fungicides. Groups with a different letter are significantly different (P<0.05).

When comparing fungicide sensitivity between baseline and exposed isolates, exposed isolates had decreased sensitivity to all three fungicides (Figure 4.3). The median percent relative growth for exposed isolates were 68.2%, 106.7%, and 67.5% for difenoconazole, fludioxonil, and pyrimethanil respectively (Figure 4.3). Exposed

101

isolates had greater percent relative growth than baseline isolates. The difference in percent relative growth between baseline and exposed isolates was statistically significant only for fludioxonil (P<0.0001) (difenoconazole P=0.57, pyrimethanil

P=0.053) (Figure 4.3). For exposed isolates, the percent relative growth ranged from

58.8–124.4%, 41.1–166.7%, and 40.5–108.5% for difenoconazole, fludioxonil, and pyrimethanil respectively. The heatmap visualizing comparisons among exposed and baseline P. niveus isolates indicated that baseline isolates grouped together and that they were more sensitive to the three fungicides as a group (Figure 4.4).

Figure 4.4. Heatmap of relationships among exposed and baseline P. niveus isolates based on percent relative growth at EC50 concentrations of difenoconazole, fludioxonil, and pyrimethanil.

102

DISCUSSION

This is the first report demonstrating that P. niveus can infect apples in the orchard. Symptoms of Paecilomyces rot on these apples included lesions and rot like those described under post-harvest conditions (Biango-Daniels and Hodge 2018). This is also the first study to assess the sensitivity of isolates of P. niveus to three fungicides with uses in both pre- and post-harvest apple production: difenoconazole, fludioxonil, and pyrimethanil. Baseline isolates were more sensitive to these fungicides than isolates from production environments where they may have been exposed to fungicides (P<0.0001). Sensitivity was highest to fludioxonil, which has no pre-harvest use in apple.

Paecilomyces (Byssochlamys) species are the most common heat-resistant molds in cider production (Swanson et al. 1985). Paecilomyces niveus is common in orchard soils, and the only member of its genus known to be a plant pathogen

(Biango-Daniels and Hodge 2018; Splittstoesser et al. 1971). Given that heat-resistant asci have been observed in infected apples, spoilage associated with this fungus could originate on infected fruit used as ingredients. Whether apples are infected post- harvest, or pre-harvest when wounds encounter P. niveus requires further investigation, as both could be possible routes to infection. Future research should also investigate P. niveus as an opportunist wound pathogen. In this study it was observed that the difference in lesion size between mock-inoculated and inoculated apples was most marked on trees with fire blight. The apple trees used from experimental block

103

three had severe fire blight symptoms, which may have contributed to the severe infection caused by P. niveus (Figure 4.2).

The external disease symptoms on orchard apples from P. niveus inoculation were similar to post-harvest symptoms described for this organism (Biango-Daniels and Hodge 2018). The lesions around the wounds were circular, with faint concentric rings of medium-brown color. The wounded faces of the inoculated apples were flattened, giving them a lopsided appearance when compared to mock-inoculated apples. This was not observed in post-harvest infections and the malformation of apples is likely the effect of infection on growth. Internal symptoms of P. niveus infection were like post-harvest symptoms, with the noteworthy difference that the necrosis, which was firm and medium brown in color, was a distorted U-shape due to the asymmetrical growth of inoculated apples. The necrotic mesocarp tissue did not easily separate from healthy tissue. Unlike post-harvest symptoms of Paecilomyces rot, there was a dark layer of tissue at the interface of the healthy and diseased tissue, which was a solid medium-brown color (Illustration 4.2 and Illustration 4.3) (Biango-

Daniels and Hodge 2018).

As P. niveus has been shown to successfully infect apples in the orchard and under post-harvest conditions, it is important to consider how to best manage this disease. Three different classes of fungicides, represented by difenoconazole, fludioxonil, and pyrimethanil, were found to have good efficacy against baseline populations of P. niveus during the mycelial growth stage, suggesting potential use in disease management. However, isolates from environments likely to have been exposed to fungicides showed reduced sensitivity to fludioxonil and pyrimethanil

104

(p<0.0001 and p=0.053, respectively). The significant reduction in sensitivity to fludioxonil among exposed isolates was surprising, as this chemical has only post- harvest uses and a low to medium fungicide resistance risk (FRAC 2018). Resistance to fludioxonil in Botrytis cinerea has been documented in berries, but is considered sporadic (Li et al. 2014). Exposed isolates of P. niveus isolates did not show significantly reduced sensitivity to pyrimethanil. Pyrimethanil carries a medium risk of fungicide resistance and remains effective against Venturia inequalis in both New

York and Chile (FRAC 2018; Henríquez et al. 2011; Köller et al. 2005).

Compared with baseline isolates, exposed isolates exhibited a wider range of relative growth values in fungicide trials. What appear to be outliers may instead reflect differences in quantitative resistance among isolates resulting from different levels of selective pressures in the environments of origin. The baseline isolates group together, with lower percent relative growth than most of the exposed isolates (Figure

4.4). Among exposed P. niveus isolates, 108-11, 146-311, and 141-1 grouped with baseline isolates according to their growth on fungicides despite their origins from orchard soils. This may indicate that reduced sensitivity, especially among exposed isolates, is not yet widespread.

Studies have shown that selection for resistant isolates can occur in the field after multiple uses of a single-site fungicide (Köller and Wilcox 2001; Villani et al.

2015). Selection for reduced-sensitivity isolates in this study may have occurred in the agricultural environments due to fungicide exposure prior to isolate collection. If P. niveus becomes more prevalent in orchards or post-harvest settings, isolates that are resistant to multiple single-site fungicides may lead to management problems. A

105

reduction in sensitivity may be due to quantitative resistance, where exposed isolates may have increased activity of efflux pumps (De Waard 1997). For example, overexpression of the ABC transporter atrB can lead to fludioxonil resistance in

Botrytis cinerea and contribute to multidrug resistance (Kretschmer et al. 2009; Li et al. 2014). This could explain why some isolates had reduced sensitivity to more than one single-site fungicide (Figure 4.4). Ultimately, confirmation of resistance requires further genotypic analysis for mutations in fungicide target sites.

If use of fungicides becomes warranted to manage P. niveus, products with difenoconazole may provide the best control out of the fungicides analyzed in this study. Currently, difenoconazole is registered in New York for use in apple to manage summer diseases such as flyspeck and sooty blotch (Cox 2017). Therefore, summer cover applications of difenoconazole could also serve to manage P. niveus along with other diseases. Careful fungicide resistance management practices should be followed if application is required to avoid any decrease in fungicide sensitivity, as difenoconazole insensitivity has been noted in other systems (Villani et al. 2015).

In summary, Paecilomyces niveus can infect developing apples in the orchard.

If the population of P. niveus from agricultural environments has already been exposed to these chemicals and responded to the selection pressure, it may be difficult for growers to manage this pathogen with fungicides containing pyrimethanil, difenoconazole, or fludioxonil. Background levels of selected quantitative resistance in P. niveus may be present and mandate higher rates of fungicides for optimal control. The development of Paecilomyces rot may be difficult to predict and will

106

remain so until more research elucidates distribution and prevalence of P. niveus in commercial orchards.

ACKNOWLEDGEMENTS

This work was supported by the USDA National Institute of Food and Agriculture,

Hatch project 1002546, and by the Cornell CALS Arthur Boller Apple Research

Grant. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the

National Institute of Food and Agriculture (NIFA) or the United States Department of

Agriculture (USDA). We thank Mei-Wah Choi for her technical support. We are grateful to Cornell Cooperative Extension agents, apple consultants, and orchard owners for their help in soil sampling. We thank the Cornell University Orchards for their support of this work, and Erika Mudrak of the Cornell Statistical Consulting Unit for her assistance with data analysis.

107

REFERENCES

Bayne, H. G., and Michener, H. D. 1979. Heat resistance of Byssochlamys ascospores.

Appl. Environ. Microbiol. 37:449–453.

Biango-Daniels, M. N., and Hodge, K. T. 2018. Paecilomyces Rot: A New Apple

Disease. Accepted for publication in Plant Disease.

https://doi.org/10.1094/PDIS-12-17-1896-RE.

Brown, A. H. S., and Smith, G. 1957. The genus Paecilomyces Bainier and its perfect

stage Byssochlamys Westling. Trans. Br. Mycol. Soc. 40:17–89.

Castresana J. 2000. Selection of conserved blocks from multiple alignments for their

use in phylogenetic analysis. Molecular Biology and Evolution. 17(4):540-552.

Cox, K.D. 2017. Pest Management Guidelines for Tree Fruit Production. “General

Pest Management Considerations.” A.M. Agnello, Coordinating & Content Ed.

Cornell University, Ithaca.

De Waard, M. A. 1997. Significance of ABC transporters in fungicide sensitivity and

resistance. Pestic. Sci. 51:271-275.

Edgar R.C. 2004. MUSCLE: multiple sequence alignment with high accuracy and

high throughput. Nucleic Acids Research. 32(5):1792-1797.

FRAC. 2018. FRAC Code List 2018: Fungicides sorted by mode of action (including

FRAC Code numbering). Available at: http://www.frac.info/docs/default-

source/publications/frac-code-list/frac_code_list_2018-

final.pdf?sfvrsn=6144b9a_2. Accessed: 25 March 2018.

108

Frederick, Z.A., Villani, S.M., Cooley, D.R., Biggs, A.R., Raes, J.J., and Cox, K.D.

2014. Prevalence and stability of qualitative QoI resistance in populations of

Venturia inaequalis in the Northeastern United States. Plant Dis. 98:1122-

1130.

Health Canada. 2016. Maximum levels for chemical contaminants in foods. Gov.

Canada. Available at: https://www.canada.ca/en/health-canada/services/food-

nutrition/food-safety/chemical-contaminants/maximum-levels-chemical-

contaminants-foods.html. Accessed: 29 January 2018.

Henríquez, J.L.S., Sarmiento, O.V. and Alarcón, P.C. 2011. Sensitivity of Venturia

inaequalis Chilean isolates to difenoconazole, fenarimol, mancozeb, and

pyrimethanil. Chilean Journal of Agricultural Research. 71(1): 39.