key: cord-0939744-xc61osdx authors: Qureshi, Abid; Thakur, Nishant; Tandon, Himani; Kumar, Manoj title: AVPdb: a database of experimentally validated antiviral peptides targeting medically important viruses date: 2014-01-01 journal: Nucleic Acids Res DOI: 10.1093/nar/gkt1191 sha: a135f7028fe8bfd36c791323bf28623c06c7b7fb doc_id: 939744 cord_uid: xc61osdx Antiviral peptides (AVPs) have exhibited huge potential in inhibiting viruses by targeting various stages of their life cycle. Therefore, we have developed AVPdb, available online at http://crdd.osdd.net/servers/avpdb, to provide a dedicated resource of experimentally verified AVPs targeting over 60 medically important viruses including Influenza, HCV, HSV, RSV, HBV, DENV, SARS, etc. However, we have separately provided HIV inhibiting peptides in ‘HIPdb’. AVPdb contains detailed information of 2683 peptides, including 624 modified peptides experimentally tested for antiviral activity. In modified peptides a chemical moiety is attached for increasing their efficacy and stability. Detailed information include: peptide sequence, length, source, virus targeted, virus family, cell line used, efficacy (qualitative/quantitative), target step/protein, assay used in determining the efficacy and PubMed reference. The database also furnishes physicochemical properties and predicted structure for each peptide. We have provided user-friendly browsing and search facility along with other analysis tools to help the users. Entering of many synthetic peptide-based drugs in various stages of clinical trials reiterate the importance for the AVP resources. AVPdb is anticipated to cater to the needs of scientific community working for the development of antiviral therapeutics. Viruses are the causative agents of various dreadful diseases in humans and animals (1, 2) . For majority of viruses like Hepatitis C virus (HCV), Influenza, Dengue virus (DENV), Severe acute respiratory syndrome (SARS), Herpes simplex virus (HSV), etc., antiviral therapeutics are limited or lacking (3) . Moreover, owing to increasing drug resistance, conventional antiviral therapy is continuously challenged these days (4, 5) . Therefore, scientific efforts are underway to search for novel antivirals (6, 7) . Antiviral peptides (AVPs) are being regarded as such new promising entities to combat the viral infections. AVPs are a subset of antimicrobial peptides (AMPs) which act as the first line of defence in many organisms as innate immune response and are the hosts' defence peptides generated in response to pathogenic disease condition (8) (9) (10) . AVPs are known to act either directly or by eliciting immune response (11) . They usually inhibit directly one or more stages in the life cycle of a virus, viz., entry, attachment, replication, transcription, translation, maturation, release, etc.; thereby exhibiting the antiviral effects (12, 13) . One of the earliest reports stating the direct involvement of peptides in inhibiting Herpes simplex virus (HSV) multiplication dates back to 1985 (14) . Since then researchers have been extensively working on peptidebased antiviral development. Bultmann et al. (15) used FGF4 signal peptide derivatives to inhibit HSV-1 entry and the best performing AVP had a half maximal inhibitory concentration (IC 50 ) of 0.7 mM. Budge and Graham (9) used r-A derived peptides to inhibit Respiratory syncytial virus (RSV) replication and achieved a maximum IC 50 of 1.23 mM. Also, a peptide derived from spike (S) protein of SARS-CoV has been proved to be effective against SARS virus entry with an efficacy of 11 mM (16). The peptide 'FluPep' inhibits Influenza virus attachment to the cells with an IC 50 of 0.10 mM (17) . Similarly, Xu et al. (18) were able to inhibit DENV protease using AVPs with a minimum IC 50 of 3.3 mM. An AVP named 'Ctry2459' has been synthesized, which possesses anti-HCV activity with an EC 50 of 1.84 mg/ml (19) . Therapeutic potential, mode of action and importance of AVPs has been further reviewed (8, 10, 20) . Peptide-based drugs are advantageous over conventional drugs in having lesser molecular weight, higher efficiency, lower toxicity and minor side effects (21) . AVPs are usually derived from natural sources but they can be readily modified by adding chemical groups or non-natural amino acids to further enhance their activity and stability (22) . Due to high potential, an estimated 15 peptidebased therapeutics as antimicrobial/immunomodulatory are under clinical trials (23) . The first AVP to pass the clinical trials was 'Enfuvirtide' (T20), an HIV fusion inhibitor that is being sold under the name of 'Fuzeon' (24, 25) . Bioinformatics resources are required to accommodate and analyse the enormous data being generated on AVPs. Although a number of resources exist for general antimicrobial peptides like APD2 (26) , CAMP (27) , DAMPD (28) , YADAMP (29) , LAMP (30), etc., yet, specific resources on AVPs are lacking. Therefore, to fill this void we have recently developed AVPpred (13) and HIPdb (12) . AVPpred is the first AVP prediction algorithm developed using Support Vector Machine (SVM). Whereas HIPdb is a specific database of experimentally validated HIV inhibiting peptides, which is freely available at http://crdd.osdd.net/servers/hipdb. HIPdb harbours information of 981 peptides and 87 modified peptides experimentally tested for HIV inhibiting activity. Besides above, no other resource is available for AVPs. Hence, we developed AVPdb-a comprehensive resource of peptides experimentally validated for their antiviral activities. Relevant data were retrieved from the PubMed database, a free repository of abstracts and references on biomedical and life sciences. Exhaustive literature search was accomplished by building search queries having combination of many keywords including virus, viral, peptide, inhibit, block, etc. A typical text mining query is given below: Full text search returned 37 842 articles as on 1 July 2013. In the initial screening, we found that majority of the articles were not furnishing the desired data. This could be due to the fact that the above keywords are quite frequent in the literature. Therefore, we limited our query to the title/abstract fields and retrieved 5000 articles using the advanced search option of PubMed. These articles were manually examined in detail based on their abstracts/full paper to fish out the desired data. Besides, we have also searched these keywords in the PatentLens database and included data from eight relevant patents. Reviews, general methodological and non-English articles were not considered. Besides these, there were number of articles in which information on only predicted peptides or peptide structures or analogues was given were excluded. Also, dendrimeric peptides, complex peptide conjugates and peptide/drug combinations were removed. Similarly, articles that were lacking peptide sequence or experimental efficacy were also not considered. In addition, peptides targeting HIV were also left out of the database, as these data were already published in our recent database, HIPdb (12) . Papers that were limited in giving information only on predicted peptides or design, peptide structural studies, peptide analogues, dendrimeric peptides, complex peptide conjugates, peptides used in combination with drugs, emphasis was laid on to articles having experimentally validated peptides and covering all or most of the AVPdb fields. After filtering out the above articles, remaining 263 research articles were finally used to collect 2059 peptides experimentally tested for virus inhibiting activity. Further 624 modified peptides were also extracted and have been provided separately in AVPdb. In our database, complete AVP data of almost all human viruses reported in the literature have been included. AVPdb is a manually curated, open source database of AVPs targeted against diverse viruses of therapeutic importance. The database comes with easy-to-operate browsing as well as searching with sorting and filtering functionalities. AVPdb also provides physicochemical properties and predicted structure of AVPs along with more informative tools for data analysis such as BLAST and MAP as well as links to major peptide resources. Physicochemical properties displayed are charge, polarity, composition, hydrophobicity and secondary structure preference. The values used for calculating these properties were retrieved from the AAindex database. Structures of the peptides were predicted using the PepStr algorithm (31) and PEP-FOLD (32) server. Structures are displayed in Jmol applet. To view the structures, java plugin should be installed in the browser and JavaScript to be enabled. BLAST and MAP tools help in finding the similar peptides reported in the database. Overall database architecture is shown in Figure 1 . AVPdb currently archives the following fields extracted from the literature: Sequence: all peptide sequences are formatted in standard one letter amino acid notation along with their respective string length. Figure 2 . These facts were also separately calculated for modified peptides as shown in Figure 3 . Further analysing the overall amino acid composition of the database, it was noticed that some amino acids like Leu, Lys, Ala, Arg and Val were found to be more abundant while some amino acids like His, Met, Trp and Tyr were present less frequently. These results are shown in Supplementary Figure S1 . Peptide efficacy statistics for natural as well as modified peptides are presented in Table 1 . Also, the top sources of the natural and modified peptides are given in Table 2 . The 'AVPdb MAP' is a user friendly tool to fetch the perfectly matching peptide available in our database. So, it helps the user to find how many peptides against the user-provided protein sequence are available in our database. The output of this tool displays the AVPid, its source, sequence and its target. Also mentioned is the start position where the match is found in the user-provided sequence. (ii) AVPdb BLAST Additionally, the BLAST allows alignment of a userprovided peptide sequence against all the peptide sequences available in our database. This helps the user to confirm whether a given peptide sequence or similar one has already been reported or not. The output is given in the standard format with the BLAST score and e-value. The alignment is shown for the peptides found to be identical or similar in the database. The output can be formatted based on the options provided by user. Various important physicochemical properties such as amino acid composition, hydrophobicity, preference for b-sheets, frequency of a-helix, amino acid charge and polarity can been calculated using AAindex (33) . These properties can be calculated for any user-provided peptide sequence by submitting it on the analysis page available under tools column. A user-friendly 'Browse by' option allows to explore the data for normal peptides by any of the fields categorized in the database, viz., Virus, Family, Peptide Source, Cell Line, Target and Assay. For modified AVPs also, a separate browse option is provided where the data can be sought by Virus, Modification, Peptide Source, Cell Line, Target and Assay. To specifically retrieve HIV inhibiting peptides, extensive links of HIPdb are provided from AVPdb pages. AVPdb has been incorporated with four different searches: (i) Field Search: here the user can enter the query in the box and can specify any of the 10 fields against which one wishes to search or else keep the default 'all' option which will search against all the fields in the database. Besides the option to choose the fields, search type allows to retrieve either an exact match or the match containing the query. The results obtained from this search display 10 fields where first nine contain the experimental data and the last one, 'Analysis', has links to BLAST results, physicochemical properties and predicted peptide structure. As more and more AVPs are being published, the interested workers may submit the desired data into AVPdb via the online submission form provided in the database. Once the information is cross-verified by our team, it will be included in the updates of database. AVPdb database is implemented using the open source LAMP solution stack on Red Hat Enterprise Linux 5 (IBM SAS Â3800 machine) with MySQL (5.0.51b) and Apache (2.2.17) in back-end and front-end of web interface is implemented with PHP (5.2.14). The database is freely available at http://crdd.osdd.net/servers/avpdb. A vast amount of data regarding AVPs both natural as well as modified is reported every year. To cope with these valuable data, we would like to include more viruses or newly discovered unique peptides to our database as appropriate information becomes available in the scientific literature. Also, a tool to predict the IC 50 value of virus inhibitory peptides shall be plugged in the database in near future. Supplementary Data are available at NAR Online. Mechanisms of viral emergence Emerging viral diseases Rates of evolutionary change in viruses: patterns and determinants Herpes simplex virus resistance to acyclovir and penciclovir after two decades of antiviral therapy Antiviral resistance and the future landscape of hepatitis C virus infection therapy HIVsirDB: a database of HIV inhibiting siRNAs VIRsiRNAdb: a curated database of experimentally validated viral siRNA/shRNA Anti herpes simplex virus activity of lactoferrin/ lactoferricin -an example of antiviral activity of antimicrobial protein/peptide Inhibition of respiratory syncytial virus by RhoA-derived peptides: implications for the development of improved antiviral agents targeting heparinbinding viruses Antimicrobial and hostdefense peptides as new anti-infective therapeutic strategies The gamma interferon (IFN-gamma) mimetic peptide IFN-gamma (95-133) prevents encephalomyocarditis virus infection both in tissue culture and in mice HIPdb: a database of experimentally validated HIV inhibiting peptides AVPpred: collection and prediction of highly effective antiviral peptides Inhibition of virus multiplication by immunoactive peptides Modified FGF4 signal peptide inhibits entry of herpes simplex virus type 1 Identification of a new region of SARS-CoV S protein critical for viral entry A novel family of peptides with potent activity against influenza A viruses Critical effect of peptide cyclization on the potency of peptide inhibitors against Dengue virus NS2B-NS3 protease Design of histidine-rich peptides with enhanced bioavailability and inhibitory activity against hepatitis C virus Antimicrobial peptides of multicellular organisms Phage display of combinatorial peptide libraries: application to antiviral research Chemical modifications designed to improve peptide stability: incorporation of non-natural amino acids, pseudo-peptide bonds, and cyclization Designing antimicrobial peptides: form follows function A phase II clinical study of the long-term safety and antiviral activity of enfuvirtide-based antiretroviral therapy Enfuvirtide (Fuzeon): the first fusion inhibitor APD2: the updated antimicrobial peptide database and its application in peptide design CAMP: a useful resource for research on antimicrobial peptides DAMPD: a manually curated antimicrobial peptide database YADAMP: yet another database of antimicrobial peptides LAMP: A database linking antimicrobial peptides PEPstr: a de novo method for tertiary structure prediction of small bioactive peptides PEP-FOLD: an updated de novo structure prediction server for both linear and disulfide bonded cyclic peptides AAindex: amino acid index database, progress report Conflict of interest statement. None declared. These peptides are comprised of non-natural or chemically modified amino acids.