AJMB Avicenna Journal of Medical Biotechnology 2008-2835 2008-4625 Avicenna Research Institute AJMB-7-45 Original Article Use of Raman Spectroscopy to Decrease Time for Identifying the Species of Candida Growth in Cultures Chouthai Nitin S. 1 * Shah Anuj A. 1 Salimnia Hossein 2 Palyvoda Olena 3 Devpura Suneetha 4 Klein Michael 5 Asmar Basim 6 Division of Neonatal-Perinatal Medicine, Wayne State University, Detroit, MI, United States of America Division of Pathology, Wayne State University, Detroit, MI, United States of America Lumigen Instrument Center, Wayne State University, Detroit, MI, United States of America Department of Radiation Oncology, Henry Ford Health System, Detroit MI, United States of America Pediatric Surgery, Wayne State University, Detroit, MI, United States of America Division of Pediatric Infectious Diseases, Wayne State University, Detroit, MI, United States of America Corresponding author: Nitin S Chouthai, MBBS, DCH, MRCP, FAAP, Division of Neonatal-Perinatal Medicine, Carman and Ann Adams Department of Pediatrics, Wayne State University Children's Hospital of Michigan and Hutzel Women's Hospital, Detroit, MI, USA. Tel: +313 745 5638. Fax: +313 745 5867. E-mail: nchoutha@med.wayne.edu Jan-Mar 2015 7 1 45 48 09 07 2014 18 10 2014 Copyright © 2015 Avicenna Research Institute 2015

This work is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported License which allows users to read, copy, distribute and make derivative works for non-commercial purposes from the material, as long as the author of the original work is cited properly.

Background

The objective of this study is to establish Raman signatures from pure cultures of different Candida species using Raman Spectroscopy (RS) and use these signatures for rapid identification of unknown Candida species.

Methods

Pure cultures of five Candida species were evaluated using RS to build a limited signature library. ‘Raman Processing’ (RP) software was used for Principal Component Analysis (PCA) and Differential Functional Analysis (DFA).

Results

Eleven principal components described at least 95% variance in the spectra. Raman signatures from these known Candida species were able to identify the species of unknown Candida cultures with 100% accuracy.

Conclusion

Raman spectroscopy can improve early identification of Candida species and may facilitate early optimal antifungal therapy.

Candida species Raman Spectrum analysis
Introduction

Late-onset sepsis occurs in 2-20% of extremely low birth weight neonates and up to 12% of these infections are due to candida species (1). Furthermore, candida Blood Stream Infections (BSIs) are associated with a 25-40% mortality rate (2). Invasive candida infections have also been associated with a 73% rate of neurodevelopmental impairment or death (3). Currently, identification of yeasts grown in a blood culture takes 24-48 hr which may delay appropriate antifungal therapy (4, 5).

Raman Spectroscopy (RS) has been used to identify many characteristics of individual cell types including cancer tissue and stem cell types (68). RS measures the scattering of light by substances that have been excited through molecular vibrations within the compound (9, 10). These molecular fingerprints can be used for qualitative identification of different compounds. It can potentially be used to create specific signatures for microorganisms like bacteria or fungi (11). Creating a reference library of specific signatures of clinically important pathogens may allow the use of spectroscopy as a tool for rapid identification of pathogens in clinical practice.

Candida albicans (C. albicans) and Candida para-psilosis (C. parapsilosis) are the most common agents that cause fungal BSIs in neonates (12). In addition, the incidence of Candida glabrata (C. glabrata) and Candida krusei (C. krusei) infections has increased due to increased use of fluconazole, to which these organisms are resistant (12, 13). A small percentage of fungal BSIs are also caused by Candida tropicalis (C. tropicalis) (12). The objective of this study was to establish signatures of C. albicans, C. parapsilosis, C. glabrata, C. krusei and C. tropicalis that have been grown on Sabouraud Dextrose Agar (SDA) using RS. A second objective was to determine if these signatures can be used to identify unknown candida isolates.

Materials and Methods

Colonies from five commonly occurring candida species grown on SDA media were evaluated. Candida species were obtained from American Type Culture Collection (C. albicans: ATCC 28815, C. parapsilosis: ATCC 34136, C. glabrata: ATCC 2001, C. krusei: ATCC 6258 and C. tropicalis: ATCC 13803). The identities of the samples were confirmed using API 20C identification system. All culture plates were inoculated using standard techniques of equal amount of inoculum and incubated for 24 hr. The plates with growth of fungi were directly placed under the Raman spectroscope without additional processing. All spectral measurements (C. albicans: 115 spectra, C. para-psilosis: 112 spectra, C. glabrata: 108 spectra, C. krusei: 117 spectra, C. tropicalis: 128 spectra from five SDA plates each and Unknown: 90 spectra from nine SDA plates) were obtained from 20-25 different points on each sample's area of interest using Renishaw Wire 2.0 software.

Raman measurements were acquired using a Reni-shaw InVia Raman microscope (Renishaw, Gloucestershire, United Kingdom) by a 785 nm (infrared, IR) excitation laser with 50×objective. A single-grating spectrograph with a 1200 line/mm grating combined with a holographic notch filter for Rayleigh scattering rejection were used. A 50× objective lens was used to focus the excitation laser beam to a spot size of ∼4 µm× 30 µm on a single colony on the SDA medium. Once the laser was initiated, collection of the back-scattered light took place with the laser power set at 100% (3 mW). Each spectrum consisted of the average of 2 collections with a 10 s collection time and an extended range of 600 to 1800 cm -1. Measurements were obtained using Renishaw Wire 2.0 software. At least 20 spectra were measured from different randomly selected points on each sample. The Raman Processing Software (RP Software) was used to import and analyze the data (14). The acquired raw spectra were corrected by subtracting background fluorescence, reducing noise, and normalizing the intensities. Data were analyzed using Principal Component Analysis (PCA) and Discriminant Function Analysis (DFA). The peaks in Raman spectra (wavenumbers) that were most significant to the classification process were identified. The biochemical correlates of these Raman peaks were also described using previously published work (15, 16).

A data bank or library of different Raman signatures for each of the five candida species was created. Unknown colonies of candida species were also derived using ATCC reference species (C. albicans, C. para-psilosis, C. glabrata, C. krusei and C. tropicalis) after creating the initial library databank. The investigating team was blinded to the identity of the unknown candida species. Colonies from cultures of unknown candida species, derived from sources other than those that were used to create the library, were compared with the known Raman signatures to identify their species.

Results

When processed with RP Software, five candida species examined showed eleven principal components generated by PCA, which account for 95.2 percent of the variance. Then the principal components were fed into DFA classifier, which enabled categorizing of unknown species. Figure 1 represents the mean normalized curves for each species and demonstrates the key differentiators among them. The chemical structure of the fungal elements represents different peaks within the Raman spectra. Based on the known spectra, the trained DFA classifier was able to identify unknown sample signatures with 100% accuracy. These differentiating peaks for different biological molecules have already been described (15, 16). The Raman shift regions (wavenumbers/cm-1) associated with significant peaks within the Raman spectra and corresponding biochemical elements are described in Table 1. The Raman signatures for all the unknown species evaluated were compiled and placed together in Figure 2. This figure again demonstrates key differences among the unknown samples. Figure 3A represents two dimensional graphical representation of Principal Component 1 and Principal Component 2, thus demonstrates first step of PCA. Similarly Figure 3B demonstrates two dimensional graphical representation of first step of DFA.

Mean normalized curves of the Raman signatures of the five candida species represented individually as well as compiled together.

Compiled mean normalized curves of Raman signatures of unknown candida species.

Two dimensional graphical representation of Principal Component Analysis (3A) and Differential Function Analysis (3B).

Wavenumbers of significant peaks in Raman spectra and corresponding biochemical elements

Wave numbers (cm -1) Biochemical elements
649 Amino acid
716 Nucleotide
910 Glucose
1003 Phenylalanine ring breathing mode
1083 Nucleic acid
1152 Protein/carotenoid
1263 Amide III band component protein
1337 Amide III/tryptophan/nucleic acid
1452 Delta CH2/Delta CH3
1603 Phenylalanine
1656 Amide bonds
Discussion

In very low birth weight neonates with fungal BSIs, it is difficult at times to obtain blood samples for cultures and the amount of blood collected is sometimes suboptimal. This limits the sensitivity of blood cultures and some neonates with candida infections might go undiagnosed until an autopsy is performed (17). There are some non-culture methods for candida identification including testing for fungal antigens and DNA available either commercially or in the evaluation phase (13, 18). Use of DNA technology for detection of candida has shown promise (13). In very low birth weight neonates, however, the detection of fungal BSIs using DNA technology is not superior to fungal lysis-centrifugation isolator system (19). The DNA technology would need use of costly reagents and technical expertise. The use of RS did not include use of any consumables and evaluation could be completed within 30 min per unknown sample.

This study is the first to demonstrate the use of RS for identification of candida species recovered in cultures. After colonies grow on solid culture media, it takes approximately 48-72 hr to identify the candida species (20). The API 20C identification system used as a standard of practice involves carbohydrate assimilation by fungal colonies. RS involves identification of peaks of chemical elements followed by comparing those using PCA and DFA techniques.

Conclusion

RS is likely to reduce the identification time to only a few minutes, since Raman signatures were useful in identifying the unknown candida species with 100% percent confidence. This would be a significant advantage for very low birth weight neonates since they have limited ability to fight the invasive fungal infections. RS appears to be useful in decreasing time of identification of specific candida species after the organism has grown from a positive culture. Further studies are needed to explore the possibility of using Raman signatures for identification of organisms directly from clinical specimens.

Acknowledgement

We acknowledge that the input is provided by Marilynn Faifax, MD, PhD (Department of Pathology, Wayne State University), Marko Jovanovski (General Surgery, Wayne State University), and Luke Reisner, PhD (Electrical and Computer Engineering, Wayne State University).

Financial Support

No financial support was received for the study.

References Stoll BJ Hansen N Fanaroff AA Wright LL Carlo WA Ehrenkranz RA Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network Pediatrics 2002 110 2 Pt 1 285 291 Greenberg RG Benjamin DK Jr Gantz MG Cotten CM Stoll BJ Walsh MC Empiric antifungal therapy and outcomes in extremely low birth weight infants with invasive candidiasis J Pediatrics 2012 161 2 264 269.e2 Benjamin DK Jr Stoll BJ Fanaroff AA McDonald SA Oh W Higgins RD Neonatal candidiasis among extremely low birth weight infants: risk factors, mortality rates, and neurodevelopmental outcomes at 18 to 22 months Pediatrics 2006 117 1 84 92 Telenti A Roberts GD Fungal blood cultures Eur J Clin Microbiol Infect Dis 1989 8 9 825 831 Matthews RC Early diagnosis of systemic candidal in-fection J Antimicrob Chemother 1993 31 6 809 812 Chan JW Taylor DS Zwerdling T Lane SM Ihara K Huser T Micro-Raman spectroscopy detects individual neoplastic and normal hematopoietic cells Biophys J 2006 90 2 648 656 Kast RE Serhatkulu GK Cao A Pandya AK Dai H Thakur JS Raman spectroscopy can differentiate malignant tumors from normal breast tissue and detect early neoplastic changes in a mouse model Biopolymers 2008 89 3 235 241 Leslie DG Kast RE Poulik JM Rabah R Sood S Auner GW Identification of pediatric brain neoplasms using Raman spectroscopy Pediatr Neurosurg 2012 48 2 109 117 Gardiner DJ Graves PR Bowley HJ Practical raman spectroscopy 1989 Berlin, New York Springer-Verlag Adar F Molecular spectroscopy workbench-Raman micro imaging-What was a concept in 1975 is now a reality Spectroscopy 2007 22 11 24 28 Ashton L Lau K Winder CL Goodacre R Raman spect-roscopy: lighting up the future of microbial identification Future Microbiol 2011 6 9 991 997 Fridkin SK Kaufman D Edwards JR Shetty S Horan T Changing incidence of Candida bloodstream infections among NICU patients in the United States:1995 -2004 Pediatrics 2006 117 5 1680 1687 Mean M Marchetti O Calandra T Bench-to-bedside review: Candida infections in the intensive care unit Critical Care 2008 12 1 204 Reisner LA Cao A Pandya AK An integrated software system for processing, analyzing, and classifying Raman spectra Chemometr Intell Lab Syst 2011 105 1 83 90 De Gelder J De Gussem K Vandenabeele P Moens L Reference database of Raman spectra of biological molecules J Raman Spectrosc 2007 38 9 1133 1147 Movasaghi Z Rehman S Rehman IU Raman spectros-copy of biological tissues Appl Spectrosc Rev 2007 42 5 493 541 Ahmad S Khan Z Mustafa AS Khan ZU Seminested PCR for diagnosis of candidemia: comparison with culture, antigen detection, and biochemical methods for species identification J Clin Microbiol 2002 40 7 2483 2489 Ahmad S Khan Z Invasive candidiasis: a review of non-culture-based laboratory diagnostic methods Indian J Med Microbiol 2012 30 3 264 269 Trovato L Betta P Romeo MG Oliveri S Detection of fungal DNA in lysis–centrifugation blood culture for the diagnosis of invasive candidiasis in neonatal patients Clin Microbiol Infect 2012 18 3 E63 E5 Morris AJ Byrne TC Madden JF Reller LB Duration of incubation of fungal cultures J Clin Microbiol 1996 34 6 1583 1585