The Utility of ITS2 Sequencing for Identifying Fungal ... · The Utility of ITS2 Sequencing for...
Transcript of The Utility of ITS2 Sequencing for Identifying Fungal ... · The Utility of ITS2 Sequencing for...
www.criver.com
EVERY STEP OF THE WAY
MICROBIAL SOLUTIONS
The Utility of ITS2 Sequencing for Identifying Fungal Contaminants In the past decade, the pharmaceutical manufacturing
industry has witnessed pronounced growth, accompanied
by an increase in oversight and new manufacturing
regulations. Industrial settings today are also faced with
pronounced prevalence of molds, a situation that can lead
to environmental monitoring excursions and suspension of
operations. Interestingly, recalls due to fungal contamination
are also on the rise.
Of the 45 recalls by the US Food and Drug Administration
(FDA) in 2012, 16% were due to mold or yeast (Table 1).
The multistate outbreak of fungal meningitis in 2012 that
claimed 48 lives was caused by product contamination
by a common mold, Exserohilum rostratum, in
methylprednisolone acetate injections. Alarmingly, this is
not the first instance; a similar outbreak of fungal infections,
including meningitis, was previously reported from a
contaminated lot of injectable methylprednisolone from a
compounding pharmacy in North Carolina. In 2006, Bausch
& Lomb made headlines when an investigation by the
United States Centers for Disease Control and Prevention
(CDC) reported an increased incidence of fungal keratitis,
which resulted in a major recall for Bausch & Lomb and
upwards of $250 million in settlements from lawsuits.
SummaryPhenotypic methods are widely
used for fungal identification.
However, identification of
these fungal isolates from
pharmaceutical environments
using standard identification
procedures is often subjective.
We have found that the use
of ITS2 gene sequencing is
substantially more accurate and
reproducible.
The Utility of ITS2 Sequencing for Identifying Fungal Contaminants
While any type of fungal contamination is undoubtedly a
cause for concern, a species-level identification is crucial to
providing a definitive root cause as part of an investigation.
Indoor air, water, personnel, and materials are the primary
sources for pharmaceutical fungal contamination. Given
the potential risk, the current good manufacturing practices
(cGMP) require that the quality of a finished pharmaceutical
product has to be assured through controlled processes.
Establishing an environmental monitoring program is
one of the most important components of an effective
manufacturing production and process control system for
aseptic production of pharmaceuticals, as well as non-
sterile products.
To aid in demonstrating compliance, many manufacturers
undertake species characterization of every environmental
isolate to provide accurate information to measure the
state of control of the manufacturing facility through the
specific detection, identification, and quantification of
microorganisms. Bacterial microorganisms found in the
manufacturing environment are frequently identified to
species level. Historically, fungal identification, especially
mold identification, has not been held to the same
standards as bacterial identification. This is primarily due
to the limitations in the existing conventional methods
in identifying fungi to the species level. Additionally, the
recognition of the pharmaceutical properties of mushrooms
established a unique place for fungal species as dietary
supplements or nutraceuticals and the identification of
fungal species has become more significant in the light of
new FDA regulations for the dietary supplement industry
(21 CFR Part 111).
Recall Date Company Product Description Reason/ProblemCompany
Status
3/20/2103 Med Prep Consulting, Inc. All compounded products Potential fungal contamination
3/17/2013 Med Prep Consulting, Inc. All compounded products Potential fungal contamination
3/16/2013 Med Prep Consulting, Inc.Magnesium sulfate 2 g in dextrose 5% in water, 50 mL for injection
Confirmed fungal contamination
10/6/2012New England Compounding Center (NECC)
Methylprednisolone acetate, betamethasone, bupivicaine, more
Sterility; linkage to meningitis outbreak
Closed
10/5/2012 Hospira, Inc. Lactated Ringer’s and 5% dextrose Container leak and mold contamination
7/13/2012 Westone Laboratories, Inc. Ear lubricant Potential contamination with pathogenic bacteria and mold
5/25/2012 Franck’s Compounding LabSterile human and veterinary prescription drugs
Presence of microorganisms and fungal growth
Closed
5/2/2012 Franck’s Compounding LabTriamcinolone acetonide P.F. 80 mg/mL
Potential fungal contamination Closed
3/31/2012 Franck’s Compounding LabTriamcinolone acetonide P.F. 80 mg/mL
Potential fungal contamination Closed
3/9/2012 Franck’s Compounding Lab Brilliant Blue G Potential fungal contamination Closed
Table 1: Recent FDA drug recalls due to fungal contamination
[email protected] • www.criver.com
Conventional Fungal Identification and Its LimitationsFungi are a diverse group of eukaryotic microorganisms
with approximately 100,000 species. The two groups of
fungi that have practical importance in the cleanroom are
molds and yeasts. Phenotypic methods are widely used
by many laboratories for fungal identification due to their
relatively lower costs. However, identification of these fungal
isolates from pharmaceutical environments using standard
identification procedures requires experienced, skilled
technologists, as the expression of the fungal phenotype
frequently depends on the media and growth conditions
that have been used. A further limitation with phenotypic
methods is the size and type of reference database used.
Many of these databases are orientated towards clinical
diagnostics and not necessarily well represented for
industrial application. Therefore, extreme care is required
while interpreting microbiological identification test results
and the trending of data. For instance, Black aspergilla,
a member of the genus Aspergillus section Nigri comprises
one of the most confusing and difficult groups to identify
due to subtle differences between the species. Aspergillus
niger, a member of the Aspergillus section Nigri, has always
been problematic in cold rooms where pharmaceutical
and biotech companies store raw materials for their
cleanroom operations. Differentiation of the members of
the A. niger aggregate complex (Table 2) has been difficult
to accomplish using morphological criteria (Samson et al.
2007).
Accurate identification is very important when out-of-
specification results are obtained, and if the contamination
source has to be determined and tracked. Remediation
efforts are not effective if inaccurate information is used
to solve a given problem. However, with the advent
of molecular identification methods, a more accurate
identification of the fungal contaminant can be obtained.
Species Conidial size (mm) Vesicle size (µm) Color and size of sclerotia (mm)
Uniseriate species
A. aculeatinus 2.5-4.5 45-80 White to cream, 0.4-0.6
A. aculeatus 3.5-5 60-80 Cream, up to 0.5
A. japonicus 3.5-5 20-35 White to cream, up to 0.5
A. uvarum 3-4 20-30 Dark brown to black, 0.5-0.8
Biseriate species
A. brasiliensis 3.5-4.5 30-45 White, 1-1.5
A. carbonarius 7-9 40-80 Pink to yellow, 1.2-1.8
A. costaricaensis 3.1-4.5 40-90 Pink to grayish yellow, 1.2-1.8
A. ellipticus 3.3-5.5 75-100 Dull yellow to brown, 0.5-1.5
A. foetidus 3.5-4.5 50-80 White, 1.2-1.8
A. heteromorphus 3.5-5 15-30 White, 0.3-0.6
A. homomorphus 5-7 50-65 -
A. ibericus 5-7 50-60 -
A. lacticoffeatus 3.4-4.1 40-65 -
A. niger 3.5-5 45-80 -
A. piperis 2.8-3.6 40-55 Yellow to pink-brown, 0.5-0.8
A. sclerotiicarbonarius 4.8-9.5 45-90 Yellow to orange to red-brown
A. sclerotioniger 4.5-6.4 30-50 Yellow to orange to red-brown
A. tubingensis 3-5 40-80 White to pink, 0.5-0.8
A. vadensis 3-4 25-35 - 1. Studies in Mycology (2007) 59:129-145.
Table 2: Morphological characteristics of different species belonging to Aspergillus section Nigri.1
Genotypic IdentificationThe use of ribosomal DNA sequences for the purposes
of organismal taxonomic classification has been in use
for many decades. More recently, ribosomal genes have
been used to study the phylogenetic relationships of fungi.
Although fungal taxonomists have been using phylogenetic
analysis to characterize, classify, and re-classify fungi
for many years, it is only recently that this approach has
gained popularity for the routine identification of fungi in
the pharmaceutical manufacturing environment. This is
primarily due to the fact that while the technology existed
for performing these analyses, applications that were
developed in a compliant manner and were able to be
validated in a cGMP laboratory have only recently been
made available through products and contract service
laboratories. The introduction of MicroSEQ®, a commercial
product available from Applied Biosystems, as well as a
number of laboratories using DNA sequencing for fungal
identification, initially provided a solution for increasing the
quality of a species-level identification for fungal organisms.
DNA sequencing provides data for identification that are
substantially more accurate and reproducible than relying
solely on visual phenotypic characteristics. This is generally
well understood and accepted. The FDA recommended the
use of genetic methods in their 2004 update to the guidance
document, “FDA Guidance for Industry. Sterile Drug
Products Produced by Aseptic Processing – Current Good
Manufacturing Practice.” In this document, the FDA states
in the section on environmental monitoring, “Genotypic
methods have been shown to be more accurate and precise
than traditional biochemical and phenotypic techniques.
These methods are especially valuable for investigations
into failures (e.g., sterility test; media fill contamination).”
Limitations with the “Out of the Box” Identification SystemDespite the acceptance and support from the scientific
community and regulatory agencies, there are still
opportunities for improvement with the current
commercially available sequence-based identification
system. The first is the coverage of the database for known
fungal species. This should not be confused with the size of
the database, as a database filled with hundreds, or even
thousands, of clinical species not encountered in the
pharmaceutical manufacturing environment adds no value
to the identification system (Rozynek et al. 2004 and Hall
et al. 2003).
A second limitation of the commercially available system is
the gene target chosen to build the DNA sequence library.
This is not a limitation of the technology, but rather a
limitation of the application. When choosing an appropriate
gene target for phylogenetic analysis, one needs to find a
target that undergoes enough genetic mutation for there to
be measurable differences in the DNA sequences of similar
but different species. However, the nucleotide differences
should not be so big that truly related species appear to
be more dissimilar than they really are. This is the great
challenge in choosing the appropriate target, and in many
cases it is simply a case of trial and error.
Limits of D2 Gene ResolutionThe D2 expansion segment of the Large Subunit (LSU) of
the ribosomal gene, as utilized in the MicroSEQ system,
typically does a good job of placing an unknown fungal
isolate into the appropriate higher level taxa (Genus, Family,
Order,) and can differentiate many species reasonably, but
not in all cases. Due to these limitations in the resolution
of the D2 segment, closely related organisms may have
identical or very similar DNA sequences. For example,
several Aspergillus niger strains were reclassified as
Aspergillus brasiliensis (Varga et al. 2007). Most significant
among the isolates was Aspergillus niger ATCC 16404 that
was reclassified as Aspergillus brasiliensis. This organism
is cited in several USP chapters as a QC organism,
including USP <61> “Microbial Limits Test – Enumeration”
and USP <71> “Sterility Test.” Because of the number of
pharmaceutical companies performing these USP tests,
it is very important to be able to correctly identify the QC
organism.
The Utility of ITS2 Sequencing for Identifying Fungal Contaminants
Unfortunately, this is another example of where virtually
all phenotypic tests, as well as D2 DNA sequencing, are
unable to differentiate these two species from one another.
Phenotypically, it has always been difficult to differentiate
between A. niger strains due to a lack of diversity in
morphological features, unstable phenotypic characters,
and the significant influence of culture conditions on the
phenotype (Rinyu et al. 1995). Furthermore, D2 sequences
provide no additional information, as the DNA sequences for
all observed A. niger and A. brasiliensis strains are identical
(Figure 1).
Merits of Using ITS2 Gene Sequencing The Internal Transcribed Spacer (ITS) regions of the
ribosomal operon have been used for fungal systematics
and classification. There are two ITS regions in the fungal
rRNA operon. The first, ITS1, is found between the 18S and
5.8S rRNA genes. The second, ITS2, is located between
the 5.8S and the 28S rRNA genes. The entire rRNA operon
is transcribed. However, after transcription, the two ITS
sequences are excised and are therefore not used for any
functional purpose. Since the ITS sequences are important
enough as spacer regions to be maintained by the cell,
but not used for any functional purpose, they accumulate
mutations at a faster rate than the 5.8S, 18S, and 28S
rRNA genes. This slightly increased rate of accumulated
mutations allows the ITS sequences to provide an improved
level of resolution as compared with the D2 sequence.
It is therefore customary among the fungal phylogeneticists
to sequence the entire stretch of ITS1-5.8S-ITS2 to use in
fungal classification. However, for the routine identification
purposes, our laboratory has found that the use of ITS2
alone is usually sufficient for species-level identification.
In support for ITS2 sequencing, Houseknecht et al.
concluded that although A. niger and A. brasiliensis are
very similar, sequencing of the entire ITS1-5.8S-ITS2 DNA
sequence shows there are five differences between the
strains of A. niger and A. brasiliensis. When comparing the
ITS2 sequence alone, there is one nucleotide difference
between the species (Figure 2); however, this difference has
been shown to be extremely reproducible and is therefore
considered to be a diagnostic indicator of the species.
Figure 1. The D2 expansion region lacks species-level resolution for the two Aspergillus spp.
N Join: 4.250 %
Aspergillus niger ATCC 16888 Type
Aspergillus brasiliensis ATCC 16404
Aspergillus brasiliensis ATCC MYA-4553 Type
Aspergillus brasiliensis NRRL 35542
Aspergillus niger NRLL 1956
Aspergillus heteromorphus
Aspergillus phoenicis
Aspergillus flavus
Aspergillus fumigatus
Figure 2. The ITS2 region clearly distinguishes the two Aspergillus spp.
N Join: 3.420 %
Aspergillus niger ATCC 16888 Type
Aspergillus brasiliensis ATCC 16404
Aspergillus brasiliensis ATCC 16404
Aspergillus brasiliensis ATCC MYA-4553 Type
Aspergillus brasiliensis NRRL 35542
Aspergillus niger NRLL 1956
Aspergillus heteromorphus
Aspergillus phoenicis
Aspergillus flavus
Aspergillus fumigatus
Aspergillus niger ATCC 16888 Type
[email protected] • www.criver.com
The Impact of the Expansion of the ITS2 Reference Database and the Frequency of Fungal ID for Customer SamplesThe accuracy of an identification is not only dependent
on the genetic target used for the identification, but
also equally dependent on the library against which the
data is compared. Having access to the most relevant,
accurate, and compliant microbial libraries is crucial to
obtaining correct species-level identifications for regulated
manufacturing operations. As a company, we recognized
early on that the libraries associated with commercial
systems were not sufficient. Since that time, we have
established the most comprehensive bacterial and fungal
sequence databases for the organisms encountered in
manufacturing environments, allowing us to provide the
highest quality identification results for unknown isolates.
The Accugenix® library generation and maintenance
activities are designed to ensure that the bacterial and
fungal reference libraries contain all possible organisms
that are relevant to the industries we serve. They are also
intended to verify that existing library entries are classified
in accordance with current literature so that tracking and
trending of environmental and objectionable organisms can
be performed efficiently. Incorrect library entries can lead to
inaccurate and inconsistent identifications and misdirected
remediation efforts.
We can demonstrate a key advantage to the continued
development of our proprietary fungal library by examining
the rates of identification of unknown organisms isolated
from pharmaceutical manufacturing facilities. We have
recently seen a dramatic increase in the number of fungal
samples submitted for genotypic identification, up to
almost 30,000 in 2012 (Figure 3). During that time, we
have quadrupled the size of the Accugenix® proprietary
ITS2 fungal identification database. The addition of novel
organisms, as well as updates in taxonomic descriptions
to the proprietary library, increases the probability of a
species-level identification. As a result of this dramatic
increase in our library database, we are now able to report
93.5% of our customer samples to a species-level ID and
the percentage of “No Match” reports has dropped from
greater than 10% to 2.6% (Figure 4).
The Utility of ITS2 Sequencing for Identifying Fungal Contaminants
Figure 3. Total number of fungal samples analyzed per year
Figure 4. Fungal library entries and the impact on customer sample identification
30000
25000
20000
15000
10000
5000
2009 2010 2011 2012
60%
50%
40%
30%
20%
10%
0%
500
600
400
300
200
100
0
90%
100%
80%
70%
2009 2010 2011 2012
New Library Entries Species
Genus No Match
ITS
2 se
quen
ce id
entif
icat
ion
(%)
Num
ber
of n
ew e
ntrie
s to
ITS
2 lib
rary
per
yea
r(2
012
tota
l ent
ries:
146
1)
Continuous maintenance requires timely evaluation of
sequences that are not identified to the species level.
Investigating groups or clusters of sequences that do not
provide a species-level identification is essential to building
a library that contains relevant organisms with a broad
range of coverage. The importance of an updated library is
further demonstrated by considering the species Penicillium
chrysogenum, a commonly occurring mold in indoor
environments. This species has gained much attention in
the pharmaceutical industry for production of the antibiotic
penicillin. However, recent phylogenetic studies showed
that the penicillin-producing strains originally described by
Alexander Fleming are not P. chrysogenum, but Penicillium
rubens (Houbraken et al. 2011). The P. rubens organism is
not present in the MicroSEQ D2 library and the phylogenetic
neighbor joining tree constructed after searching the D2
database with the P. rubens sequence is disjointed and
provides inaccurate information (Figure 5).
It is critical that the identification libraries against which
data are compared contain the most current taxonomic
classifications, have a breadth of coverage and contain
relevant organisms — those that have been identified
as a result of EM programs. If the library lacks depth of
coverage, the interpretation of the data may not always
be reliable. Over the last 20 years, the Charles River
Accugenix® group has identified more than one million
microorganisms primarily isolated from manufacturing
environments. As such, we understand the biodiversity
present in these environments. We know the frequency
with which organisms are recovered and the variety of
organisms that are isolated from manufacturing facilities.
We continually strive to improve our genotypic reference
databases to improve the frequency of species-level
identification by encompassing the organisms that are
relevant to the industries we serve. Our goal is to always
provide the highest level of accuracy and reliability.
N Join: 0.375 %
Penicillum rubens CBS 205.57
Specimen Penicillum rubens CBS 205.57
MicroSeqID Phylogenetic TreeITS2 Phylogenetic Tree
Penicillum chrysogenum
Penicillum melanoconidum
Penicillum glandicola
Penicillum rubens
Penicillum verrucosum
Penicillum commune
Penicillum venetum
Penicillum turbatum
Penicillum griseofulvum
Penicillum cyclopium
Penicillum camembertii CBS 190.67
Penicillum griseofulvum DSM 896
Penicillum aurantiogriseum aurantiogriseum CBS 324.89
Penicillum commune CBS 474.92
Penicillum histrum DSM 62833
Penicillum hordei CBS 701.68
Penicillum nalgiovense CBS 352.48
Penicillum verrucosum verrucosum CBS 603.74
Penicillum rubens CBS 205.57
Penicillum chrysogenum CBS 306.48
Penicillum viridicatum DSM 2447
N Join: 1.0 %
Figure 5. Impact of updated libraries and species identification
[email protected] • www.criver.com
[email protected] • www.criver.com © 2017, Charles River Laboratories International, Inc.
ReferencesExophiala infection from contaminated injectable steroids
prepared by a compounding pharmacy -- United States.
MMWR Morb Mortal Wkly Rep 2002; 51:1109-1112.
Hall L, Wohlfiel S, Roberts GD. Experience with the
MicroSEQ D2 Large-Subunit Ribosomal DNA Sequencing
Kit for Identification of Filamentous Fungi Encountered
in the Clinical Laboratory. J Clinical Microbiology 2003;
42(2):622-626.
Houbraken J, Frisvad JC, Samson RA. Fleming’s penicillin
producing strain is not Penicillium chrysogenum but
P. rubens. IMA Fungus 2011; 2(1):87-95.
Houseknecht J, Stamenova E, Suh S-O, Beck B, McKee M,
Zhou J. Reclassification of ATCC® 16404™ and ATCC®
9642™ as Aspergillus brasiliensis. Pharm Micro Forum
Newsletter 2008; 14(10):2-8.
Rinyu E, Varga J, Ferenczy L. Phenotypic and Genotypic
Analysis of Variability in Aspergillus fumigatus. J Clinical
Microbiology 1995; 33(10):2567-2575.
Rozynek P, Gilges S, Brüning T, Wilhelm M. Quality Test
of the MicroSEQ D2 LSU Fungal Sequencing Kit for the
Identification of Fungi. Int J Hyg Environ Health 2004;
207(3):297-299.
Samson RA, Noonim P, Meijer M, Houbraken J, Frisvad JC,
Varga J. Diagnostic tools to identify black Aspergilli. Studies
in Mycology 2007; 59:129–145.
Varga J, Kocsube S, Toth B, Frisvad JC, Perrone G, Susca
A, Meiijjer M, Samson R. Aspergillus brasiliensis sp. Nov.,
a biseriate black Aspergillus species with world-wide
distribution. Int J Sys Evol Microbiology 2007;
57:1925-1932.